CHANGES revision 216166
155714Skris 255714Skris OpenSSL CHANGES 355714Skris _______________ 455714Skris 5216166Ssimon Changes between 0.9.8p and 0.9.8q [2 Dec 2010] 6216166Ssimon 7216166Ssimon *) Disable code workaround for ancient and obsolete Netscape browsers 8216166Ssimon and servers: an attacker can use it in a ciphersuite downgrade attack. 9216166Ssimon Thanks to Martin Rex for discovering this bug. CVE-2010-4180 10216166Ssimon [Steve Henson] 11216166Ssimon 12216166Ssimon *) Fixed J-PAKE implementation error, originally discovered by 13216166Ssimon Sebastien Martini, further info and confirmation from Stefan 14216166Ssimon Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252 15216166Ssimon [Ben Laurie] 16216166Ssimon 17215697Ssimon Changes between 0.9.8o and 0.9.8p [16 Nov 2010] 18215697Ssimon 19215697Ssimon *) Fix extension code to avoid race conditions which can result in a buffer 20215697Ssimon overrun vulnerability: resumed sessions must not be modified as they can 21215697Ssimon be shared by multiple threads. CVE-2010-3864 22215697Ssimon [Steve Henson] 23215697Ssimon 24215697Ssimon *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939 25215697Ssimon [Steve Henson] 26215697Ssimon 27215697Ssimon *) Don't reencode certificate when calculating signature: cache and use 28215697Ssimon the original encoding instead. This makes signature verification of 29215697Ssimon some broken encodings work correctly. 30215697Ssimon [Steve Henson] 31215697Ssimon 32215697Ssimon *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT 33215697Ssimon is also one of the inputs. 34215697Ssimon [Emilia K�sper <emilia.kasper@esat.kuleuven.be> (Google)] 35215697Ssimon 36215697Ssimon *) Don't repeatedly append PBE algorithms to table if they already exist. 37215697Ssimon Sort table on each new add. This effectively makes the table read only 38215697Ssimon after all algorithms are added and subsequent calls to PKCS12_pbe_add 39215697Ssimon etc are non-op. 40215697Ssimon [Steve Henson] 41215697Ssimon 42215697Ssimon Changes between 0.9.8n and 0.9.8o [01 Jun 2010] 43215697Ssimon 44215697Ssimon [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after 45215697Ssimon OpenSSL 1.0.0.] 46215697Ssimon 47215697Ssimon *) Correct a typo in the CMS ASN1 module which can result in invalid memory 48215697Ssimon access or freeing data twice (CVE-2010-0742) 49215697Ssimon [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>] 50215697Ssimon 51215697Ssimon *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more 52215697Ssimon common in certificates and some applications which only call 53215697Ssimon SSL_library_init and not OpenSSL_add_all_algorithms() will fail. 54215697Ssimon [Steve Henson] 55215697Ssimon 56215697Ssimon *) VMS fixes: 57215697Ssimon Reduce copying into .apps and .test in makevms.com 58215697Ssimon Don't try to use blank CA certificate in CA.com 59215697Ssimon Allow use of C files from original directories in maketests.com 60215697Ssimon [Steven M. Schweda" <sms@antinode.info>] 61215697Ssimon 62206046Ssimon Changes between 0.9.8m and 0.9.8n [24 Mar 2010] 63206046Ssimon 64206046Ssimon *) When rejecting SSL/TLS records due to an incorrect version number, never 65206046Ssimon update s->server with a new major version number. As of 66206046Ssimon - OpenSSL 0.9.8m if 'short' is a 16-bit type, 67206046Ssimon - OpenSSL 0.9.8f if 'short' is longer than 16 bits, 68206046Ssimon the previous behavior could result in a read attempt at NULL when 69206046Ssimon receiving specific incorrect SSL/TLS records once record payload 70206046Ssimon protection is active. (CVE-2010-0740) 71206046Ssimon [Bodo Moeller, Adam Langley <agl@chromium.org>] 72206046Ssimon 73206046Ssimon *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL 74206046Ssimon could be crashed if the relevant tables were not present (e.g. chrooted). 75206046Ssimon [Tomas Hoger <thoger@redhat.com>] 76206046Ssimon 77205128Ssimon Changes between 0.9.8l and 0.9.8m [25 Feb 2010] 78205128Ssimon 79205128Ssimon *) Always check bn_wexpend() return values for failure. (CVE-2009-3245) 80205128Ssimon [Martin Olsson, Neel Mehta] 81205128Ssimon 82205128Ssimon *) Fix X509_STORE locking: Every 'objs' access requires a lock (to 83205128Ssimon accommodate for stack sorting, always a write lock!). 84205128Ssimon [Bodo Moeller] 85205128Ssimon 86205128Ssimon *) On some versions of WIN32 Heap32Next is very slow. This can cause 87205128Ssimon excessive delays in the RAND_poll(): over a minute. As a workaround 88205128Ssimon include a time check in the inner Heap32Next loop too. 89205128Ssimon [Steve Henson] 90205128Ssimon 91205128Ssimon *) The code that handled flushing of data in SSL/TLS originally used the 92205128Ssimon BIO_CTRL_INFO ctrl to see if any data was pending first. This caused 93205128Ssimon the problem outlined in PR#1949. The fix suggested there however can 94205128Ssimon trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions 95205128Ssimon of Apache). So instead simplify the code to flush unconditionally. 96205128Ssimon This should be fine since flushing with no data to flush is a no op. 97205128Ssimon [Steve Henson] 98205128Ssimon 99205128Ssimon *) Handle TLS versions 2.0 and later properly and correctly use the 100205128Ssimon highest version of TLS/SSL supported. Although TLS >= 2.0 is some way 101205128Ssimon off ancient servers have a habit of sticking around for a while... 102205128Ssimon [Steve Henson] 103205128Ssimon 104205128Ssimon *) Modify compression code so it frees up structures without using the 105205128Ssimon ex_data callbacks. This works around a problem where some applications 106205128Ssimon call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when 107205128Ssimon restarting) then use compression (e.g. SSL with compression) later. 108205128Ssimon This results in significant per-connection memory leaks and 109205128Ssimon has caused some security issues including CVE-2008-1678 and 110205128Ssimon CVE-2009-4355. 111205128Ssimon [Steve Henson] 112205128Ssimon 113205128Ssimon *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't 114205128Ssimon change when encrypting or decrypting. 115205128Ssimon [Bodo Moeller] 116205128Ssimon 117205128Ssimon *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to 118205128Ssimon connect and renegotiate with servers which do not support RI. 119205128Ssimon Until RI is more widely deployed this option is enabled by default. 120205128Ssimon [Steve Henson] 121205128Ssimon 122205128Ssimon *) Add "missing" ssl ctrls to clear options and mode. 123205128Ssimon [Steve Henson] 124205128Ssimon 125205128Ssimon *) If client attempts to renegotiate and doesn't support RI respond with 126205128Ssimon a no_renegotiation alert as required by RFC5746. Some renegotiating 127205128Ssimon TLS clients will continue a connection gracefully when they receive 128205128Ssimon the alert. Unfortunately OpenSSL mishandled this alert and would hang 129205128Ssimon waiting for a server hello which it will never receive. Now we treat a 130205128Ssimon received no_renegotiation alert as a fatal error. This is because 131205128Ssimon applications requesting a renegotiation might well expect it to succeed 132205128Ssimon and would have no code in place to handle the server denying it so the 133205128Ssimon only safe thing to do is to terminate the connection. 134205128Ssimon [Steve Henson] 135205128Ssimon 136205128Ssimon *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if 137205128Ssimon peer supports secure renegotiation and 0 otherwise. Print out peer 138205128Ssimon renegotiation support in s_client/s_server. 139205128Ssimon [Steve Henson] 140205128Ssimon 141205128Ssimon *) Replace the highly broken and deprecated SPKAC certification method with 142205128Ssimon the updated NID creation version. This should correctly handle UTF8. 143205128Ssimon [Steve Henson] 144205128Ssimon 145205128Ssimon *) Implement RFC5746. Re-enable renegotiation but require the extension 146205128Ssimon as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION 147205128Ssimon turns out to be a bad idea. It has been replaced by 148205128Ssimon SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with 149205128Ssimon SSL_CTX_set_options(). This is really not recommended unless you 150205128Ssimon know what you are doing. 151205128Ssimon [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson] 152205128Ssimon 153205128Ssimon *) Fixes to stateless session resumption handling. Use initial_ctx when 154205128Ssimon issuing and attempting to decrypt tickets in case it has changed during 155205128Ssimon servername handling. Use a non-zero length session ID when attempting 156205128Ssimon stateless session resumption: this makes it possible to determine if 157205128Ssimon a resumption has occurred immediately after receiving server hello 158205128Ssimon (several places in OpenSSL subtly assume this) instead of later in 159205128Ssimon the handshake. 160205128Ssimon [Steve Henson] 161205128Ssimon 162205128Ssimon *) The functions ENGINE_ctrl(), OPENSSL_isservice(), 163205128Ssimon CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error 164205128Ssimon fixes for a few places where the return code is not checked 165205128Ssimon correctly. 166205128Ssimon [Julia Lawall <julia@diku.dk>] 167205128Ssimon 168205128Ssimon *) Add --strict-warnings option to Configure script to include devteam 169205128Ssimon warnings in other configurations. 170205128Ssimon [Steve Henson] 171205128Ssimon 172205128Ssimon *) Add support for --libdir option and LIBDIR variable in makefiles. This 173205128Ssimon makes it possible to install openssl libraries in locations which 174205128Ssimon have names other than "lib", for example "/usr/lib64" which some 175205128Ssimon systems need. 176205128Ssimon [Steve Henson, based on patch from Jeremy Utley] 177205128Ssimon 178205128Ssimon *) Don't allow the use of leading 0x80 in OIDs. This is a violation of 179205128Ssimon X690 8.9.12 and can produce some misleading textual output of OIDs. 180205128Ssimon [Steve Henson, reported by Dan Kaminsky] 181205128Ssimon 182205128Ssimon *) Delete MD2 from algorithm tables. This follows the recommendation in 183205128Ssimon several standards that it is not used in new applications due to 184205128Ssimon several cryptographic weaknesses. For binary compatibility reasons 185205128Ssimon the MD2 API is still compiled in by default. 186205128Ssimon [Steve Henson] 187205128Ssimon 188205128Ssimon *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved 189205128Ssimon and restored. 190205128Ssimon [Steve Henson] 191205128Ssimon 192205128Ssimon *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and 193205128Ssimon OPENSSL_asc2uni conditionally on Netware platforms to avoid a name 194205128Ssimon clash. 195205128Ssimon [Guenter <lists@gknw.net>] 196205128Ssimon 197205128Ssimon *) Fix the server certificate chain building code to use X509_verify_cert(), 198205128Ssimon it used to have an ad-hoc builder which was unable to cope with anything 199205128Ssimon other than a simple chain. 200205128Ssimon [David Woodhouse <dwmw2@infradead.org>, Steve Henson] 201205128Ssimon 202205128Ssimon *) Don't check self signed certificate signatures in X509_verify_cert() 203205128Ssimon by default (a flag can override this): it just wastes time without 204205128Ssimon adding any security. As a useful side effect self signed root CAs 205205128Ssimon with non-FIPS digests are now usable in FIPS mode. 206205128Ssimon [Steve Henson] 207205128Ssimon 208205128Ssimon *) In dtls1_process_out_of_seq_message() the check if the current message 209205128Ssimon is already buffered was missing. For every new message was memory 210205128Ssimon allocated, allowing an attacker to perform an denial of service attack 211205128Ssimon with sending out of seq handshake messages until there is no memory 212205128Ssimon left. Additionally every future messege was buffered, even if the 213205128Ssimon sequence number made no sense and would be part of another handshake. 214205128Ssimon So only messages with sequence numbers less than 10 in advance will be 215205128Ssimon buffered. (CVE-2009-1378) 216205128Ssimon [Robin Seggelmann, discovered by Daniel Mentz] 217205128Ssimon 218205128Ssimon *) Records are buffered if they arrive with a future epoch to be 219205128Ssimon processed after finishing the corresponding handshake. There is 220205128Ssimon currently no limitation to this buffer allowing an attacker to perform 221205128Ssimon a DOS attack with sending records with future epochs until there is no 222205128Ssimon memory left. This patch adds the pqueue_size() function to detemine 223205128Ssimon the size of a buffer and limits the record buffer to 100 entries. 224205128Ssimon (CVE-2009-1377) 225205128Ssimon [Robin Seggelmann, discovered by Daniel Mentz] 226205128Ssimon 227205128Ssimon *) Keep a copy of frag->msg_header.frag_len so it can be used after the 228205128Ssimon parent structure is freed. (CVE-2009-1379) 229205128Ssimon [Daniel Mentz] 230205128Ssimon 231205128Ssimon *) Handle non-blocking I/O properly in SSL_shutdown() call. 232205128Ssimon [Darryl Miles <darryl-mailinglists@netbauds.net>] 233205128Ssimon 234205128Ssimon *) Add 2.5.4.* OIDs 235205128Ssimon [Ilya O. <vrghost@gmail.com>] 236205128Ssimon 237205128Ssimon Changes between 0.9.8k and 0.9.8l [5 Nov 2009] 238205128Ssimon 239205128Ssimon *) Disable renegotiation completely - this fixes a severe security 240205128Ssimon problem (CVE-2009-3555) at the cost of breaking all 241205128Ssimon renegotiation. Renegotiation can be re-enabled by setting 242205128Ssimon SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at 243205128Ssimon run-time. This is really not recommended unless you know what 244205128Ssimon you're doing. 245205128Ssimon [Ben Laurie] 246205128Ssimon 247194206Ssimon Changes between 0.9.8j and 0.9.8k [25 Mar 2009] 248167612Ssimon 249194206Ssimon *) Don't set val to NULL when freeing up structures, it is freed up by 250194206Ssimon underlying code. If sizeof(void *) > sizeof(long) this can result in 251194206Ssimon zeroing past the valid field. (CVE-2009-0789) 252194206Ssimon [Paolo Ganci <Paolo.Ganci@AdNovum.CH>] 253194206Ssimon 254194206Ssimon *) Fix bug where return value of CMS_SignerInfo_verify_content() was not 255194206Ssimon checked correctly. This would allow some invalid signed attributes to 256194206Ssimon appear to verify correctly. (CVE-2009-0591) 257194206Ssimon [Ivan Nestlerode <inestlerode@us.ibm.com>] 258194206Ssimon 259194206Ssimon *) Reject UniversalString and BMPString types with invalid lengths. This 260194206Ssimon prevents a crash in ASN1_STRING_print_ex() which assumes the strings have 261194206Ssimon a legal length. (CVE-2009-0590) 262194206Ssimon [Steve Henson] 263194206Ssimon 264194206Ssimon *) Set S/MIME signing as the default purpose rather than setting it 265194206Ssimon unconditionally. This allows applications to override it at the store 266194206Ssimon level. 267194206Ssimon [Steve Henson] 268194206Ssimon 269194206Ssimon *) Permit restricted recursion of ASN1 strings. This is needed in practice 270194206Ssimon to handle some structures. 271194206Ssimon [Steve Henson] 272194206Ssimon 273194206Ssimon *) Improve efficiency of mem_gets: don't search whole buffer each time 274194206Ssimon for a '\n' 275194206Ssimon [Jeremy Shapiro <jnshapir@us.ibm.com>] 276194206Ssimon 277194206Ssimon *) New -hex option for openssl rand. 278194206Ssimon [Matthieu Herrb] 279194206Ssimon 280194206Ssimon *) Print out UTF8String and NumericString when parsing ASN1. 281194206Ssimon [Steve Henson] 282194206Ssimon 283194206Ssimon *) Support NumericString type for name components. 284194206Ssimon [Steve Henson] 285194206Ssimon 286194206Ssimon *) Allow CC in the environment to override the automatically chosen 287194206Ssimon compiler. Note that nothing is done to ensure flags work with the 288194206Ssimon chosen compiler. 289194206Ssimon [Ben Laurie] 290194206Ssimon 291194206Ssimon Changes between 0.9.8i and 0.9.8j [07 Jan 2009] 292194206Ssimon 293194206Ssimon *) Properly check EVP_VerifyFinal() and similar return values 294194206Ssimon (CVE-2008-5077). 295194206Ssimon [Ben Laurie, Bodo Moeller, Google Security Team] 296194206Ssimon 297194206Ssimon *) Enable TLS extensions by default. 298194206Ssimon [Ben Laurie] 299194206Ssimon 300194206Ssimon *) Allow the CHIL engine to be loaded, whether the application is 301194206Ssimon multithreaded or not. (This does not release the developer from the 302194206Ssimon obligation to set up the dynamic locking callbacks.) 303194206Ssimon [Sander Temme <sander@temme.net>] 304194206Ssimon 305194206Ssimon *) Use correct exit code if there is an error in dgst command. 306194206Ssimon [Steve Henson; problem pointed out by Roland Dirlewanger] 307194206Ssimon 308194206Ssimon *) Tweak Configure so that you need to say "experimental-jpake" to enable 309194206Ssimon JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications. 310194206Ssimon [Bodo Moeller] 311194206Ssimon 312194206Ssimon *) Add experimental JPAKE support, including demo authentication in 313194206Ssimon s_client and s_server. 314194206Ssimon [Ben Laurie] 315194206Ssimon 316194206Ssimon *) Set the comparison function in v3_addr_canonize(). 317194206Ssimon [Rob Austein <sra@hactrn.net>] 318194206Ssimon 319194206Ssimon *) Add support for XMPP STARTTLS in s_client. 320194206Ssimon [Philip Paeps <philip@freebsd.org>] 321194206Ssimon 322194206Ssimon *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior 323194206Ssimon to ensure that even with this option, only ciphersuites in the 324194206Ssimon server's preference list will be accepted. (Note that the option 325194206Ssimon applies only when resuming a session, so the earlier behavior was 326194206Ssimon just about the algorithm choice for symmetric cryptography.) 327194206Ssimon [Bodo Moeller] 328194206Ssimon 329194206Ssimon Changes between 0.9.8h and 0.9.8i [15 Sep 2008] 330194206Ssimon 331205128Ssimon *) Fix NULL pointer dereference if a DTLS server received 332205128Ssimon ChangeCipherSpec as first record (CVE-2009-1386). 333205128Ssimon [PR #1679] 334205128Ssimon 335194206Ssimon *) Fix a state transitition in s3_srvr.c and d1_srvr.c 336194206Ssimon (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...). 337194206Ssimon [Nagendra Modadugu] 338194206Ssimon 339194206Ssimon *) The fix in 0.9.8c that supposedly got rid of unsafe 340194206Ssimon double-checked locking was incomplete for RSA blinding, 341194206Ssimon addressing just one layer of what turns out to have been 342194206Ssimon doubly unsafe triple-checked locking. 343194206Ssimon 344194206Ssimon So now fix this for real by retiring the MONT_HELPER macro 345194206Ssimon in crypto/rsa/rsa_eay.c. 346194206Ssimon 347194206Ssimon [Bodo Moeller; problem pointed out by Marius Schilder] 348194206Ssimon 349194206Ssimon *) Various precautionary measures: 350194206Ssimon 351194206Ssimon - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h). 352194206Ssimon 353194206Ssimon - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c). 354194206Ssimon (NB: This would require knowledge of the secret session ticket key 355194206Ssimon to exploit, in which case you'd be SOL either way.) 356194206Ssimon 357194206Ssimon - Change bn_nist.c so that it will properly handle input BIGNUMs 358194206Ssimon outside the expected range. 359194206Ssimon 360194206Ssimon - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG 361194206Ssimon builds. 362194206Ssimon 363194206Ssimon [Neel Mehta, Bodo Moeller] 364194206Ssimon 365194206Ssimon *) Allow engines to be "soft loaded" - i.e. optionally don't die if 366194206Ssimon the load fails. Useful for distros. 367194206Ssimon [Ben Laurie and the FreeBSD team] 368194206Ssimon 369194206Ssimon *) Add support for Local Machine Keyset attribute in PKCS#12 files. 370194206Ssimon [Steve Henson] 371194206Ssimon 372194206Ssimon *) Fix BN_GF2m_mod_arr() top-bit cleanup code. 373194206Ssimon [Huang Ying] 374194206Ssimon 375194206Ssimon *) Expand ENGINE to support engine supplied SSL client certificate functions. 376194206Ssimon 377194206Ssimon This work was sponsored by Logica. 378194206Ssimon [Steve Henson] 379194206Ssimon 380194206Ssimon *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows 381194206Ssimon keystores. Support for SSL/TLS client authentication too. 382194206Ssimon Not compiled unless enable-capieng specified to Configure. 383194206Ssimon 384194206Ssimon This work was sponsored by Logica. 385194206Ssimon [Steve Henson] 386194206Ssimon 387194206Ssimon *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using 388194206Ssimon ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain 389194206Ssimon attribute creation routines such as certifcate requests and PKCS#12 390194206Ssimon files. 391194206Ssimon [Steve Henson] 392194206Ssimon 393194206Ssimon Changes between 0.9.8g and 0.9.8h [28 May 2008] 394194206Ssimon 395194206Ssimon *) Fix flaw if 'Server Key exchange message' is omitted from a TLS 396194206Ssimon handshake which could lead to a cilent crash as found using the 397194206Ssimon Codenomicon TLS test suite (CVE-2008-1672) 398194206Ssimon [Steve Henson, Mark Cox] 399194206Ssimon 400194206Ssimon *) Fix double free in TLS server name extensions which could lead to 401194206Ssimon a remote crash found by Codenomicon TLS test suite (CVE-2008-0891) 402194206Ssimon [Joe Orton] 403194206Ssimon 404194206Ssimon *) Clear error queue in SSL_CTX_use_certificate_chain_file() 405194206Ssimon 406194206Ssimon Clear the error queue to ensure that error entries left from 407194206Ssimon older function calls do not interfere with the correct operation. 408194206Ssimon [Lutz Jaenicke, Erik de Castro Lopo] 409194206Ssimon 410194206Ssimon *) Remove root CA certificates of commercial CAs: 411194206Ssimon 412194206Ssimon The OpenSSL project does not recommend any specific CA and does not 413194206Ssimon have any policy with respect to including or excluding any CA. 414194206Ssimon Therefore it does not make any sense to ship an arbitrary selection 415194206Ssimon of root CA certificates with the OpenSSL software. 416194206Ssimon [Lutz Jaenicke] 417194206Ssimon 418194206Ssimon *) RSA OAEP patches to fix two separate invalid memory reads. 419194206Ssimon The first one involves inputs when 'lzero' is greater than 420194206Ssimon 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes 421194206Ssimon before the beginning of from). The second one involves inputs where 422194206Ssimon the 'db' section contains nothing but zeroes (there is a one-byte 423194206Ssimon invalid read after the end of 'db'). 424194206Ssimon [Ivan Nestlerode <inestlerode@us.ibm.com>] 425194206Ssimon 426194206Ssimon *) Partial backport from 0.9.9-dev: 427194206Ssimon 428194206Ssimon Introduce bn_mul_mont (dedicated Montgomery multiplication 429194206Ssimon procedure) as a candidate for BIGNUM assembler implementation. 430194206Ssimon While 0.9.9-dev uses assembler for various architectures, only 431194206Ssimon x86_64 is available by default here in the 0.9.8 branch, and 432194206Ssimon 32-bit x86 is available through a compile-time setting. 433194206Ssimon 434194206Ssimon To try the 32-bit x86 assembler implementation, use Configure 435194206Ssimon option "enable-montasm" (which exists only for this backport). 436194206Ssimon 437194206Ssimon As "enable-montasm" for 32-bit x86 disclaims code stability 438194206Ssimon anyway, in this constellation we activate additional code 439194206Ssimon backported from 0.9.9-dev for further performance improvements, 440194206Ssimon namely BN_from_montgomery_word. (To enable this otherwise, 441194206Ssimon e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".) 442194206Ssimon 443194206Ssimon [Andy Polyakov (backport partially by Bodo Moeller)] 444194206Ssimon 445194206Ssimon *) Add TLS session ticket callback. This allows an application to set 446194206Ssimon TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed 447194206Ssimon values. This is useful for key rollover for example where several key 448194206Ssimon sets may exist with different names. 449194206Ssimon [Steve Henson] 450194206Ssimon 451194206Ssimon *) Reverse ENGINE-internal logic for caching default ENGINE handles. 452194206Ssimon This was broken until now in 0.9.8 releases, such that the only way 453194206Ssimon a registered ENGINE could be used (assuming it initialises 454194206Ssimon successfully on the host) was to explicitly set it as the default 455194206Ssimon for the relevant algorithms. This is in contradiction with 0.9.7 456194206Ssimon behaviour and the documentation. With this fix, when an ENGINE is 457194206Ssimon registered into a given algorithm's table of implementations, the 458194206Ssimon 'uptodate' flag is reset so that auto-discovery will be used next 459194206Ssimon time a new context for that algorithm attempts to select an 460194206Ssimon implementation. 461194206Ssimon [Ian Lister (tweaked by Geoff Thorpe)] 462194206Ssimon 463194206Ssimon *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9 464194206Ssimon implemention in the following ways: 465194206Ssimon 466194206Ssimon Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be 467194206Ssimon hard coded. 468194206Ssimon 469194206Ssimon Lack of BER streaming support means one pass streaming processing is 470194206Ssimon only supported if data is detached: setting the streaming flag is 471194206Ssimon ignored for embedded content. 472194206Ssimon 473194206Ssimon CMS support is disabled by default and must be explicitly enabled 474194206Ssimon with the enable-cms configuration option. 475194206Ssimon [Steve Henson] 476194206Ssimon 477194206Ssimon *) Update the GMP engine glue to do direct copies between BIGNUM and 478194206Ssimon mpz_t when openssl and GMP use the same limb size. Otherwise the 479194206Ssimon existing "conversion via a text string export" trick is still used. 480194206Ssimon [Paul Sheer <paulsheer@gmail.com>] 481194206Ssimon 482194206Ssimon *) Zlib compression BIO. This is a filter BIO which compressed and 483194206Ssimon uncompresses any data passed through it. 484194206Ssimon [Steve Henson] 485194206Ssimon 486194206Ssimon *) Add AES_wrap_key() and AES_unwrap_key() functions to implement 487194206Ssimon RFC3394 compatible AES key wrapping. 488194206Ssimon [Steve Henson] 489194206Ssimon 490194206Ssimon *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0(): 491194206Ssimon sets string data without copying. X509_ALGOR_set0() and 492194206Ssimon X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier) 493194206Ssimon data. Attribute function X509at_get0_data_by_OBJ(): retrieves data 494194206Ssimon from an X509_ATTRIBUTE structure optionally checking it occurs only 495194206Ssimon once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied 496194206Ssimon data. 497194206Ssimon [Steve Henson] 498194206Ssimon 499194206Ssimon *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set() 500194206Ssimon to get the expected BN_FLG_CONSTTIME behavior. 501194206Ssimon [Bodo Moeller (Google)] 502194206Ssimon 503194206Ssimon *) Netware support: 504194206Ssimon 505194206Ssimon - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets 506194206Ssimon - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT) 507194206Ssimon - added some more tests to do_tests.pl 508194206Ssimon - fixed RunningProcess usage so that it works with newer LIBC NDKs too 509194206Ssimon - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency 510194206Ssimon - added new Configure targets netware-clib-bsdsock, netware-clib-gcc, 511194206Ssimon netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc 512194206Ssimon - various changes to netware.pl to enable gcc-cross builds on Win32 513194206Ssimon platform 514194206Ssimon - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD) 515194206Ssimon - various changes to fix missing prototype warnings 516194206Ssimon - fixed x86nasm.pl to create correct asm files for NASM COFF output 517194206Ssimon - added AES, WHIRLPOOL and CPUID assembler code to build files 518194206Ssimon - added missing AES assembler make rules to mk1mf.pl 519194206Ssimon - fixed order of includes in apps/ocsp.c so that e_os.h settings apply 520194206Ssimon [Guenter Knauf <eflash@gmx.net>] 521194206Ssimon 522194206Ssimon *) Implement certificate status request TLS extension defined in RFC3546. 523194206Ssimon A client can set the appropriate parameters and receive the encoded 524194206Ssimon OCSP response via a callback. A server can query the supplied parameters 525194206Ssimon and set the encoded OCSP response in the callback. Add simplified examples 526194206Ssimon to s_client and s_server. 527194206Ssimon [Steve Henson] 528194206Ssimon 529194206Ssimon Changes between 0.9.8f and 0.9.8g [19 Oct 2007] 530194206Ssimon 531194206Ssimon *) Fix various bugs: 532194206Ssimon + Binary incompatibility of ssl_ctx_st structure 533194206Ssimon + DTLS interoperation with non-compliant servers 534194206Ssimon + Don't call get_session_cb() without proposed session 535194206Ssimon + Fix ia64 assembler code 536194206Ssimon [Andy Polyakov, Steve Henson] 537194206Ssimon 538194206Ssimon Changes between 0.9.8e and 0.9.8f [11 Oct 2007] 539194206Ssimon 540194206Ssimon *) DTLS Handshake overhaul. There were longstanding issues with 541194206Ssimon OpenSSL DTLS implementation, which were making it impossible for 542194206Ssimon RFC 4347 compliant client to communicate with OpenSSL server. 543194206Ssimon Unfortunately just fixing these incompatibilities would "cut off" 544194206Ssimon pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e 545194206Ssimon server keeps tolerating non RFC compliant syntax. The opposite is 546194206Ssimon not true, 0.9.8f client can not communicate with earlier server. 547194206Ssimon This update even addresses CVE-2007-4995. 548194206Ssimon [Andy Polyakov] 549194206Ssimon 550194206Ssimon *) Changes to avoid need for function casts in OpenSSL: some compilers 551194206Ssimon (gcc 4.2 and later) reject their use. 552194206Ssimon [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>, 553194206Ssimon Steve Henson] 554194206Ssimon 555194206Ssimon *) Add RFC4507 support to OpenSSL. This includes the corrections in 556194206Ssimon RFC4507bis. The encrypted ticket format is an encrypted encoded 557194206Ssimon SSL_SESSION structure, that way new session features are automatically 558194206Ssimon supported. 559194206Ssimon 560194206Ssimon If a client application caches session in an SSL_SESSION structure 561194206Ssimon support is transparent because tickets are now stored in the encoded 562194206Ssimon SSL_SESSION. 563194206Ssimon 564194206Ssimon The SSL_CTX structure automatically generates keys for ticket 565194206Ssimon protection in servers so again support should be possible 566194206Ssimon with no application modification. 567194206Ssimon 568194206Ssimon If a client or server wishes to disable RFC4507 support then the option 569194206Ssimon SSL_OP_NO_TICKET can be set. 570194206Ssimon 571194206Ssimon Add a TLS extension debugging callback to allow the contents of any client 572194206Ssimon or server extensions to be examined. 573194206Ssimon 574194206Ssimon This work was sponsored by Google. 575194206Ssimon [Steve Henson] 576194206Ssimon 577194206Ssimon *) Add initial support for TLS extensions, specifically for the server_name 578194206Ssimon extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 579194206Ssimon have new members for a host name. The SSL data structure has an 580194206Ssimon additional member SSL_CTX *initial_ctx so that new sessions can be 581194206Ssimon stored in that context to allow for session resumption, even after the 582194206Ssimon SSL has been switched to a new SSL_CTX in reaction to a client's 583194206Ssimon server_name extension. 584194206Ssimon 585194206Ssimon New functions (subject to change): 586194206Ssimon 587194206Ssimon SSL_get_servername() 588194206Ssimon SSL_get_servername_type() 589194206Ssimon SSL_set_SSL_CTX() 590194206Ssimon 591194206Ssimon New CTRL codes and macros (subject to change): 592194206Ssimon 593194206Ssimon SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 594194206Ssimon - SSL_CTX_set_tlsext_servername_callback() 595194206Ssimon SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 596194206Ssimon - SSL_CTX_set_tlsext_servername_arg() 597194206Ssimon SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 598194206Ssimon 599194206Ssimon openssl s_client has a new '-servername ...' option. 600194206Ssimon 601194206Ssimon openssl s_server has new options '-servername_host ...', '-cert2 ...', 602194206Ssimon '-key2 ...', '-servername_fatal' (subject to change). This allows 603194206Ssimon testing the HostName extension for a specific single host name ('-cert' 604194206Ssimon and '-key' remain fallbacks for handshakes without HostName 605194206Ssimon negotiation). If the unrecogninzed_name alert has to be sent, this by 606194206Ssimon default is a warning; it becomes fatal with the '-servername_fatal' 607194206Ssimon option. 608194206Ssimon 609194206Ssimon [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson] 610194206Ssimon 611194206Ssimon *) Add AES and SSE2 assembly language support to VC++ build. 612194206Ssimon [Steve Henson] 613194206Ssimon 614194206Ssimon *) Mitigate attack on final subtraction in Montgomery reduction. 615194206Ssimon [Andy Polyakov] 616194206Ssimon 617194206Ssimon *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0 618194206Ssimon (which previously caused an internal error). 619194206Ssimon [Bodo Moeller] 620194206Ssimon 621194206Ssimon *) Squeeze another 10% out of IGE mode when in != out. 622194206Ssimon [Ben Laurie] 623194206Ssimon 624194206Ssimon *) AES IGE mode speedup. 625194206Ssimon [Dean Gaudet (Google)] 626194206Ssimon 627194206Ssimon *) Add the Korean symmetric 128-bit cipher SEED (see 628194206Ssimon http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and 629194206Ssimon add SEED ciphersuites from RFC 4162: 630194206Ssimon 631194206Ssimon TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA" 632194206Ssimon TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA" 633194206Ssimon TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA" 634194206Ssimon TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA" 635194206Ssimon 636194206Ssimon To minimize changes between patchlevels in the OpenSSL 0.9.8 637194206Ssimon series, SEED remains excluded from compilation unless OpenSSL 638194206Ssimon is configured with 'enable-seed'. 639194206Ssimon [KISA, Bodo Moeller] 640194206Ssimon 641194206Ssimon *) Mitigate branch prediction attacks, which can be practical if a 642194206Ssimon single processor is shared, allowing a spy process to extract 643194206Ssimon information. For detailed background information, see 644194206Ssimon http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron, 645194206Ssimon J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL 646194206Ssimon and Necessary Software Countermeasures"). The core of the change 647194206Ssimon are new versions BN_div_no_branch() and 648194206Ssimon BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(), 649194206Ssimon respectively, which are slower, but avoid the security-relevant 650194206Ssimon conditional branches. These are automatically called by BN_div() 651194206Ssimon and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one 652194206Ssimon of the input BIGNUMs. Also, BN_is_bit_set() has been changed to 653194206Ssimon remove a conditional branch. 654194206Ssimon 655194206Ssimon BN_FLG_CONSTTIME is the new name for the previous 656194206Ssimon BN_FLG_EXP_CONSTTIME flag, since it now affects more than just 657194206Ssimon modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag 658194206Ssimon in the exponent causes BN_mod_exp_mont() to use the alternative 659194206Ssimon implementation in BN_mod_exp_mont_consttime().) The old name 660194206Ssimon remains as a deprecated alias. 661194206Ssimon 662194206Ssimon Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general 663194206Ssimon RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses 664194206Ssimon constant-time implementations for more than just exponentiation. 665194206Ssimon Here too the old name is kept as a deprecated alias. 666194206Ssimon 667194206Ssimon BN_BLINDING_new() will now use BN_dup() for the modulus so that 668194206Ssimon the BN_BLINDING structure gets an independent copy of the 669194206Ssimon modulus. This means that the previous "BIGNUM *m" argument to 670194206Ssimon BN_BLINDING_new() and to BN_BLINDING_create_param() now 671194206Ssimon essentially becomes "const BIGNUM *m", although we can't actually 672194206Ssimon change this in the header file before 0.9.9. It allows 673194206Ssimon RSA_setup_blinding() to use BN_with_flags() on the modulus to 674194206Ssimon enable BN_FLG_CONSTTIME. 675194206Ssimon 676194206Ssimon [Matthew D Wood (Intel Corp)] 677194206Ssimon 678194206Ssimon *) In the SSL/TLS server implementation, be strict about session ID 679194206Ssimon context matching (which matters if an application uses a single 680194206Ssimon external cache for different purposes). Previously, 681194206Ssimon out-of-context reuse was forbidden only if SSL_VERIFY_PEER was 682194206Ssimon set. This did ensure strict client verification, but meant that, 683194206Ssimon with applications using a single external cache for quite 684194206Ssimon different requirements, clients could circumvent ciphersuite 685194206Ssimon restrictions for a given session ID context by starting a session 686194206Ssimon in a different context. 687194206Ssimon [Bodo Moeller] 688194206Ssimon 689167612Ssimon *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 690167612Ssimon a ciphersuite string such as "DEFAULT:RSA" cannot enable 691167612Ssimon authentication-only ciphersuites. 692167612Ssimon [Bodo Moeller] 693167612Ssimon 694194206Ssimon *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was 695194206Ssimon not complete and could lead to a possible single byte overflow 696194206Ssimon (CVE-2007-5135) [Ben Laurie] 697194206Ssimon 698194206Ssimon Changes between 0.9.8d and 0.9.8e [23 Feb 2007] 699194206Ssimon 700167612Ssimon *) Since AES128 and AES256 (and similarly Camellia128 and 701167612Ssimon Camellia256) share a single mask bit in the logic of 702167612Ssimon ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 703167612Ssimon kludge to work properly if AES128 is available and AES256 isn't 704167612Ssimon (or if Camellia128 is available and Camellia256 isn't). 705167612Ssimon [Victor Duchovni] 706167612Ssimon 707167612Ssimon *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c 708167612Ssimon (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters): 709167612Ssimon When a point or a seed is encoded in a BIT STRING, we need to 710167612Ssimon prevent the removal of trailing zero bits to get the proper DER 711167612Ssimon encoding. (By default, crypto/asn1/a_bitstr.c assumes the case 712167612Ssimon of a NamedBitList, for which trailing 0 bits need to be removed.) 713167612Ssimon [Bodo Moeller] 714167612Ssimon 715167612Ssimon *) Have SSL/TLS server implementation tolerate "mismatched" record 716167612Ssimon protocol version while receiving ClientHello even if the 717167612Ssimon ClientHello is fragmented. (The server can't insist on the 718167612Ssimon particular protocol version it has chosen before the ServerHello 719167612Ssimon message has informed the client about his choice.) 720167612Ssimon [Bodo Moeller] 721167612Ssimon 722167612Ssimon *) Add RFC 3779 support. 723167612Ssimon [Rob Austein for ARIN, Ben Laurie] 724167612Ssimon 725167612Ssimon *) Load error codes if they are not already present instead of using a 726167612Ssimon static variable. This allows them to be cleanly unloaded and reloaded. 727167612Ssimon Improve header file function name parsing. 728167612Ssimon [Steve Henson] 729167612Ssimon 730167612Ssimon *) extend SMTP and IMAP protocol emulation in s_client to use EHLO 731167612Ssimon or CAPABILITY handshake as required by RFCs. 732167612Ssimon [Goetz Babin-Ebell] 733167612Ssimon 734162911Ssimon Changes between 0.9.8c and 0.9.8d [28 Sep 2006] 735162911Ssimon 736162911Ssimon *) Introduce limits to prevent malicious keys being able to 737162911Ssimon cause a denial of service. (CVE-2006-2940) 738162911Ssimon [Steve Henson, Bodo Moeller] 739162911Ssimon 740162911Ssimon *) Fix ASN.1 parsing of certain invalid structures that can result 741162911Ssimon in a denial of service. (CVE-2006-2937) [Steve Henson] 742162911Ssimon 743162911Ssimon *) Fix buffer overflow in SSL_get_shared_ciphers() function. 744162911Ssimon (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 745162911Ssimon 746162911Ssimon *) Fix SSL client code which could crash if connecting to a 747162911Ssimon malicious SSLv2 server. (CVE-2006-4343) 748162911Ssimon [Tavis Ormandy and Will Drewry, Google Security Team] 749162911Ssimon 750162911Ssimon *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites 751162911Ssimon match only those. Before that, "AES256-SHA" would be interpreted 752162911Ssimon as a pattern and match "AES128-SHA" too (since AES128-SHA got 753162911Ssimon the same strength classification in 0.9.7h) as we currently only 754162911Ssimon have a single AES bit in the ciphersuite description bitmap. 755162911Ssimon That change, however, also applied to ciphersuite strings such as 756162911Ssimon "RC4-MD5" that intentionally matched multiple ciphersuites -- 757162911Ssimon namely, SSL 2.0 ciphersuites in addition to the more common ones 758162911Ssimon from SSL 3.0/TLS 1.0. 759162911Ssimon 760162911Ssimon So we change the selection algorithm again: Naming an explicit 761162911Ssimon ciphersuite selects this one ciphersuite, and any other similar 762162911Ssimon ciphersuite (same bitmap) from *other* protocol versions. 763162911Ssimon Thus, "RC4-MD5" again will properly select both the SSL 2.0 764162911Ssimon ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite. 765162911Ssimon 766162911Ssimon Since SSL 2.0 does not have any ciphersuites for which the 767162911Ssimon 128/256 bit distinction would be relevant, this works for now. 768162911Ssimon The proper fix will be to use different bits for AES128 and 769162911Ssimon AES256, which would have avoided the problems from the beginning; 770162911Ssimon however, bits are scarce, so we can only do this in a new release 771162911Ssimon (not just a patchlevel) when we can change the SSL_CIPHER 772162911Ssimon definition to split the single 'unsigned long mask' bitmap into 773162911Ssimon multiple values to extend the available space. 774162911Ssimon 775162911Ssimon [Bodo Moeller] 776162911Ssimon 777162911Ssimon Changes between 0.9.8b and 0.9.8c [05 Sep 2006] 778162911Ssimon 779162911Ssimon *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 780162911Ssimon (CVE-2006-4339) [Ben Laurie and Google Security Team] 781162911Ssimon 782162911Ssimon *) Add AES IGE and biIGE modes. 783162911Ssimon [Ben Laurie] 784162911Ssimon 785162911Ssimon *) Change the Unix randomness entropy gathering to use poll() when 786162911Ssimon possible instead of select(), since the latter has some 787162911Ssimon undesirable limitations. 788162911Ssimon [Darryl Miles via Richard Levitte and Bodo Moeller] 789162911Ssimon 790162911Ssimon *) Disable "ECCdraft" ciphersuites more thoroughly. Now special 791162911Ssimon treatment in ssl/ssl_ciph.s makes sure that these ciphersuites 792162911Ssimon cannot be implicitly activated as part of, e.g., the "AES" alias. 793162911Ssimon However, please upgrade to OpenSSL 0.9.9[-dev] for 794162911Ssimon non-experimental use of the ECC ciphersuites to get TLS extension 795162911Ssimon support, which is required for curve and point format negotiation 796162911Ssimon to avoid potential handshake problems. 797162911Ssimon [Bodo Moeller] 798162911Ssimon 799162911Ssimon *) Disable rogue ciphersuites: 800162911Ssimon 801162911Ssimon - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 802162911Ssimon - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 803162911Ssimon - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 804162911Ssimon 805162911Ssimon The latter two were purportedly from 806162911Ssimon draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 807162911Ssimon appear there. 808162911Ssimon 809167612Ssimon Also deactivate the remaining ciphersuites from 810162911Ssimon draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 811162911Ssimon unofficial, and the ID has long expired. 812162911Ssimon [Bodo Moeller] 813162911Ssimon 814162911Ssimon *) Fix RSA blinding Heisenbug (problems sometimes occured on 815162911Ssimon dual-core machines) and other potential thread-safety issues. 816162911Ssimon [Bodo Moeller] 817162911Ssimon 818162911Ssimon *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key 819162911Ssimon versions), which is now available for royalty-free use 820162911Ssimon (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html). 821162911Ssimon Also, add Camellia TLS ciphersuites from RFC 4132. 822162911Ssimon 823162911Ssimon To minimize changes between patchlevels in the OpenSSL 0.9.8 824162911Ssimon series, Camellia remains excluded from compilation unless OpenSSL 825162911Ssimon is configured with 'enable-camellia'. 826162911Ssimon [NTT] 827162911Ssimon 828162911Ssimon *) Disable the padding bug check when compression is in use. The padding 829162911Ssimon bug check assumes the first packet is of even length, this is not 830162911Ssimon necessarily true if compresssion is enabled and can result in false 831162911Ssimon positives causing handshake failure. The actual bug test is ancient 832162911Ssimon code so it is hoped that implementations will either have fixed it by 833162911Ssimon now or any which still have the bug do not support compression. 834162911Ssimon [Steve Henson] 835162911Ssimon 836160814Ssimon Changes between 0.9.8a and 0.9.8b [04 May 2006] 837160814Ssimon 838160814Ssimon *) When applying a cipher rule check to see if string match is an explicit 839160814Ssimon cipher suite and only match that one cipher suite if it is. 840160814Ssimon [Steve Henson] 841160814Ssimon 842160814Ssimon *) Link in manifests for VC++ if needed. 843160814Ssimon [Austin Ziegler <halostatue@gmail.com>] 844160814Ssimon 845160814Ssimon *) Update support for ECC-based TLS ciphersuites according to 846160814Ssimon draft-ietf-tls-ecc-12.txt with proposed changes (but without 847160814Ssimon TLS extensions, which are supported starting with the 0.9.9 848160814Ssimon branch, not in the OpenSSL 0.9.8 branch). 849160814Ssimon [Douglas Stebila] 850160814Ssimon 851160814Ssimon *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support 852160814Ssimon opaque EVP_CIPHER_CTX handling. 853160814Ssimon [Steve Henson] 854160814Ssimon 855160814Ssimon *) Fixes and enhancements to zlib compression code. We now only use 856160814Ssimon "zlib1.dll" and use the default __cdecl calling convention on Win32 857160814Ssimon to conform with the standards mentioned here: 858160814Ssimon http://www.zlib.net/DLL_FAQ.txt 859160814Ssimon Static zlib linking now works on Windows and the new --with-zlib-include 860160814Ssimon --with-zlib-lib options to Configure can be used to supply the location 861160814Ssimon of the headers and library. Gracefully handle case where zlib library 862160814Ssimon can't be loaded. 863160814Ssimon [Steve Henson] 864160814Ssimon 865160814Ssimon *) Several fixes and enhancements to the OID generation code. The old code 866160814Ssimon sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't 867160814Ssimon handle numbers larger than ULONG_MAX, truncated printing and had a 868160814Ssimon non standard OBJ_obj2txt() behaviour. 869160814Ssimon [Steve Henson] 870160814Ssimon 871160814Ssimon *) Add support for building of engines under engine/ as shared libraries 872160814Ssimon under VC++ build system. 873160814Ssimon [Steve Henson] 874160814Ssimon 875160814Ssimon *) Corrected the numerous bugs in the Win32 path splitter in DSO. 876160814Ssimon Hopefully, we will not see any false combination of paths any more. 877160814Ssimon [Richard Levitte] 878160814Ssimon 879160814Ssimon Changes between 0.9.8 and 0.9.8a [11 Oct 2005] 880160814Ssimon 881160814Ssimon *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 882160814Ssimon (part of SSL_OP_ALL). This option used to disable the 883160814Ssimon countermeasure against man-in-the-middle protocol-version 884160814Ssimon rollback in the SSL 2.0 server implementation, which is a bad 885160814Ssimon idea. (CVE-2005-2969) 886160814Ssimon 887160814Ssimon [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 888160814Ssimon for Information Security, National Institute of Advanced Industrial 889160814Ssimon Science and Technology [AIST], Japan)] 890160814Ssimon 891160814Ssimon *) Add two function to clear and return the verify parameter flags. 892160814Ssimon [Steve Henson] 893160814Ssimon 894160814Ssimon *) Keep cipherlists sorted in the source instead of sorting them at 895160814Ssimon runtime, thus removing the need for a lock. 896160814Ssimon [Nils Larsch] 897160814Ssimon 898160814Ssimon *) Avoid some small subgroup attacks in Diffie-Hellman. 899160814Ssimon [Nick Mathewson and Ben Laurie] 900160814Ssimon 901160814Ssimon *) Add functions for well-known primes. 902160814Ssimon [Nick Mathewson] 903160814Ssimon 904160814Ssimon *) Extended Windows CE support. 905160814Ssimon [Satoshi Nakamura and Andy Polyakov] 906160814Ssimon 907160814Ssimon *) Initialize SSL_METHOD structures at compile time instead of during 908160814Ssimon runtime, thus removing the need for a lock. 909160814Ssimon [Steve Henson] 910160814Ssimon 911160814Ssimon *) Make PKCS7_decrypt() work even if no certificate is supplied by 912160814Ssimon attempting to decrypt each encrypted key in turn. Add support to 913160814Ssimon smime utility. 914160814Ssimon [Steve Henson] 915160814Ssimon 916160814Ssimon Changes between 0.9.7h and 0.9.8 [05 Jul 2005] 917160814Ssimon 918162911Ssimon [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after 919162911Ssimon OpenSSL 0.9.8.] 920162911Ssimon 921160814Ssimon *) Add libcrypto.pc and libssl.pc for those who feel they need them. 922160814Ssimon [Richard Levitte] 923160814Ssimon 924160814Ssimon *) Change CA.sh and CA.pl so they don't bundle the CSR and the private 925160814Ssimon key into the same file any more. 926160814Ssimon [Richard Levitte] 927160814Ssimon 928160814Ssimon *) Add initial support for Win64, both IA64 and AMD64/x64 flavors. 929160814Ssimon [Andy Polyakov] 930160814Ssimon 931160814Ssimon *) Add -utf8 command line and config file option to 'ca'. 932160814Ssimon [Stefan <stf@udoma.org] 933160814Ssimon 934160814Ssimon *) Removed the macro des_crypt(), as it seems to conflict with some 935160814Ssimon libraries. Use DES_crypt(). 936160814Ssimon [Richard Levitte] 937160814Ssimon 938160814Ssimon *) Correct naming of the 'chil' and '4758cca' ENGINEs. This 939160814Ssimon involves renaming the source and generated shared-libs for 940160814Ssimon both. The engines will accept the corrected or legacy ids 941160814Ssimon ('ncipher' and '4758_cca' respectively) when binding. NB, 942160814Ssimon this only applies when building 'shared'. 943160814Ssimon [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe] 944160814Ssimon 945160814Ssimon *) Add attribute functions to EVP_PKEY structure. Modify 946160814Ssimon PKCS12_create() to recognize a CSP name attribute and 947160814Ssimon use it. Make -CSP option work again in pkcs12 utility. 948160814Ssimon [Steve Henson] 949160814Ssimon 950160814Ssimon *) Add new functionality to the bn blinding code: 951160814Ssimon - automatic re-creation of the BN_BLINDING parameters after 952160814Ssimon a fixed number of uses (currently 32) 953160814Ssimon - add new function for parameter creation 954160814Ssimon - introduce flags to control the update behaviour of the 955160814Ssimon BN_BLINDING parameters 956160814Ssimon - hide BN_BLINDING structure 957160814Ssimon Add a second BN_BLINDING slot to the RSA structure to improve 958160814Ssimon performance when a single RSA object is shared among several 959160814Ssimon threads. 960160814Ssimon [Nils Larsch] 961160814Ssimon 962160814Ssimon *) Add support for DTLS. 963160814Ssimon [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie] 964160814Ssimon 965160814Ssimon *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1) 966160814Ssimon to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file() 967160814Ssimon [Walter Goulet] 968160814Ssimon 969160814Ssimon *) Remove buggy and incompletet DH cert support from 970160814Ssimon ssl/ssl_rsa.c and ssl/s3_both.c 971160814Ssimon [Nils Larsch] 972160814Ssimon 973160814Ssimon *) Use SHA-1 instead of MD5 as the default digest algorithm for 974160814Ssimon the apps/openssl applications. 975160814Ssimon [Nils Larsch] 976160814Ssimon 977160814Ssimon *) Compile clean with "-Wall -Wmissing-prototypes 978160814Ssimon -Wstrict-prototypes -Wmissing-declarations -Werror". Currently 979160814Ssimon DEBUG_SAFESTACK must also be set. 980160814Ssimon [Ben Laurie] 981160814Ssimon 982160814Ssimon *) Change ./Configure so that certain algorithms can be disabled by default. 983160814Ssimon The new counterpiece to "no-xxx" is "enable-xxx". 984160814Ssimon 985160814Ssimon The patented RC5 and MDC2 algorithms will now be disabled unless 986160814Ssimon "enable-rc5" and "enable-mdc2", respectively, are specified. 987160814Ssimon 988160814Ssimon (IDEA remains enabled despite being patented. This is because IDEA 989160814Ssimon is frequently required for interoperability, and there is no license 990160814Ssimon fee for non-commercial use. As before, "no-idea" can be used to 991160814Ssimon avoid this algorithm.) 992160814Ssimon 993160814Ssimon [Bodo Moeller] 994160814Ssimon 995160814Ssimon *) Add processing of proxy certificates (see RFC 3820). This work was 996160814Ssimon sponsored by KTH (The Royal Institute of Technology in Stockholm) and 997160814Ssimon EGEE (Enabling Grids for E-science in Europe). 998160814Ssimon [Richard Levitte] 999160814Ssimon 1000160814Ssimon *) RC4 performance overhaul on modern architectures/implementations, such 1001160814Ssimon as Intel P4, IA-64 and AMD64. 1002160814Ssimon [Andy Polyakov] 1003160814Ssimon 1004160814Ssimon *) New utility extract-section.pl. This can be used specify an alternative 1005160814Ssimon section number in a pod file instead of having to treat each file as 1006160814Ssimon a separate case in Makefile. This can be done by adding two lines to the 1007160814Ssimon pod file: 1008160814Ssimon 1009160814Ssimon =for comment openssl_section:XXX 1010160814Ssimon 1011160814Ssimon The blank line is mandatory. 1012160814Ssimon 1013160814Ssimon [Steve Henson] 1014160814Ssimon 1015160814Ssimon *) New arguments -certform, -keyform and -pass for s_client and s_server 1016160814Ssimon to allow alternative format key and certificate files and passphrase 1017160814Ssimon sources. 1018160814Ssimon [Steve Henson] 1019160814Ssimon 1020160814Ssimon *) New structure X509_VERIFY_PARAM which combines current verify parameters, 1021160814Ssimon update associated structures and add various utility functions. 1022160814Ssimon 1023160814Ssimon Add new policy related verify parameters, include policy checking in 1024160814Ssimon standard verify code. Enhance 'smime' application with extra parameters 1025160814Ssimon to support policy checking and print out. 1026160814Ssimon [Steve Henson] 1027160814Ssimon 1028160814Ssimon *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3 1029160814Ssimon Nehemiah processors. These extensions support AES encryption in hardware 1030160814Ssimon as well as RNG (though RNG support is currently disabled). 1031160814Ssimon [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov] 1032160814Ssimon 1033160814Ssimon *) Deprecate BN_[get|set]_params() functions (they were ignored internally). 1034160814Ssimon [Geoff Thorpe] 1035160814Ssimon 1036160814Ssimon *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented. 1037160814Ssimon [Andy Polyakov and a number of other people] 1038160814Ssimon 1039160814Ssimon *) Improved PowerPC platform support. Most notably BIGNUM assembler 1040160814Ssimon implementation contributed by IBM. 1041160814Ssimon [Suresh Chari, Peter Waltenberg, Andy Polyakov] 1042160814Ssimon 1043160814Ssimon *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public 1044160814Ssimon exponent rather than 'unsigned long'. There is a corresponding change to 1045160814Ssimon the new 'rsa_keygen' element of the RSA_METHOD structure. 1046160814Ssimon [Jelte Jansen, Geoff Thorpe] 1047160814Ssimon 1048160814Ssimon *) Functionality for creating the initial serial number file is now 1049160814Ssimon moved from CA.pl to the 'ca' utility with a new option -create_serial. 1050160814Ssimon 1051160814Ssimon (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial 1052160814Ssimon number file to 1, which is bound to cause problems. To avoid 1053160814Ssimon the problems while respecting compatibility between different 0.9.7 1054160814Ssimon patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in 1055160814Ssimon CA.pl for serial number initialization. With the new release 0.9.8, 1056160814Ssimon we can fix the problem directly in the 'ca' utility.) 1057160814Ssimon [Steve Henson] 1058160814Ssimon 1059160814Ssimon *) Reduced header interdepencies by declaring more opaque objects in 1060160814Ssimon ossl_typ.h. As a consequence, including some headers (eg. engine.h) will 1061160814Ssimon give fewer recursive includes, which could break lazy source code - so 1062160814Ssimon this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always, 1063160814Ssimon developers should define this symbol when building and using openssl to 1064160814Ssimon ensure they track the recommended behaviour, interfaces, [etc], but 1065160814Ssimon backwards-compatible behaviour prevails when this isn't defined. 1066160814Ssimon [Geoff Thorpe] 1067160814Ssimon 1068160814Ssimon *) New function X509_POLICY_NODE_print() which prints out policy nodes. 1069160814Ssimon [Steve Henson] 1070160814Ssimon 1071160814Ssimon *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality. 1072160814Ssimon This will generate a random key of the appropriate length based on the 1073160814Ssimon cipher context. The EVP_CIPHER can provide its own random key generation 1074160814Ssimon routine to support keys of a specific form. This is used in the des and 1075160814Ssimon 3des routines to generate a key of the correct parity. Update S/MIME 1076160814Ssimon code to use new functions and hence generate correct parity DES keys. 1077160814Ssimon Add EVP_CHECK_DES_KEY #define to return an error if the key is not 1078160814Ssimon valid (weak or incorrect parity). 1079160814Ssimon [Steve Henson] 1080160814Ssimon 1081160814Ssimon *) Add a local set of CRLs that can be used by X509_verify_cert() as well 1082160814Ssimon as looking them up. This is useful when the verified structure may contain 1083160814Ssimon CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs 1084160814Ssimon present unless the new PKCS7_NO_CRL flag is asserted. 1085160814Ssimon [Steve Henson] 1086160814Ssimon 1087160814Ssimon *) Extend ASN1 oid configuration module. It now additionally accepts the 1088160814Ssimon syntax: 1089160814Ssimon 1090160814Ssimon shortName = some long name, 1.2.3.4 1091160814Ssimon [Steve Henson] 1092160814Ssimon 1093160814Ssimon *) Reimplemented the BN_CTX implementation. There is now no more static 1094160814Ssimon limitation on the number of variables it can handle nor the depth of the 1095160814Ssimon "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack 1096160814Ssimon information can now expand as required, and rather than having a single 1097160814Ssimon static array of bignums, BN_CTX now uses a linked-list of such arrays 1098160814Ssimon allowing it to expand on demand whilst maintaining the usefulness of 1099160814Ssimon BN_CTX's "bundling". 1100160814Ssimon [Geoff Thorpe] 1101160814Ssimon 1102160814Ssimon *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD 1103160814Ssimon to allow all RSA operations to function using a single BN_CTX. 1104160814Ssimon [Geoff Thorpe] 1105160814Ssimon 1106160814Ssimon *) Preliminary support for certificate policy evaluation and checking. This 1107160814Ssimon is initially intended to pass the tests outlined in "Conformance Testing 1108160814Ssimon of Relying Party Client Certificate Path Processing Logic" v1.07. 1109160814Ssimon [Steve Henson] 1110160814Ssimon 1111160814Ssimon *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and 1112160814Ssimon remained unused and not that useful. A variety of other little bignum 1113160814Ssimon tweaks and fixes have also been made continuing on from the audit (see 1114160814Ssimon below). 1115160814Ssimon [Geoff Thorpe] 1116160814Ssimon 1117160814Ssimon *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with 1118160814Ssimon associated ASN1, EVP and SSL functions and old ASN1 macros. 1119160814Ssimon [Richard Levitte] 1120160814Ssimon 1121160814Ssimon *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results, 1122160814Ssimon and this should never fail. So the return value from the use of 1123160814Ssimon BN_set_word() (which can fail due to needless expansion) is now deprecated; 1124160814Ssimon if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro. 1125160814Ssimon [Geoff Thorpe] 1126160814Ssimon 1127160814Ssimon *) BN_CTX_get() should return zero-valued bignums, providing the same 1128160814Ssimon initialised value as BN_new(). 1129160814Ssimon [Geoff Thorpe, suggested by Ulf M�ller] 1130160814Ssimon 1131160814Ssimon *) Support for inhibitAnyPolicy certificate extension. 1132160814Ssimon [Steve Henson] 1133160814Ssimon 1134160814Ssimon *) An audit of the BIGNUM code is underway, for which debugging code is 1135160814Ssimon enabled when BN_DEBUG is defined. This makes stricter enforcements on what 1136160814Ssimon is considered valid when processing BIGNUMs, and causes execution to 1137160814Ssimon assert() when a problem is discovered. If BN_DEBUG_RAND is defined, 1138160814Ssimon further steps are taken to deliberately pollute unused data in BIGNUM 1139160814Ssimon structures to try and expose faulty code further on. For now, openssl will 1140160814Ssimon (in its default mode of operation) continue to tolerate the inconsistent 1141160814Ssimon forms that it has tolerated in the past, but authors and packagers should 1142160814Ssimon consider trying openssl and their own applications when compiled with 1143160814Ssimon these debugging symbols defined. It will help highlight potential bugs in 1144160814Ssimon their own code, and will improve the test coverage for OpenSSL itself. At 1145160814Ssimon some point, these tighter rules will become openssl's default to improve 1146160814Ssimon maintainability, though the assert()s and other overheads will remain only 1147160814Ssimon in debugging configurations. See bn.h for more details. 1148160814Ssimon [Geoff Thorpe, Nils Larsch, Ulf M�ller] 1149160814Ssimon 1150160814Ssimon *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure 1151160814Ssimon that can only be obtained through BN_CTX_new() (which implicitly 1152160814Ssimon initialises it). The presence of this function only made it possible 1153160814Ssimon to overwrite an existing structure (and cause memory leaks). 1154160814Ssimon [Geoff Thorpe] 1155160814Ssimon 1156160814Ssimon *) Because of the callback-based approach for implementing LHASH as a 1157160814Ssimon template type, lh_insert() adds opaque objects to hash-tables and 1158160814Ssimon lh_doall() or lh_doall_arg() are typically used with a destructor callback 1159160814Ssimon to clean up those corresponding objects before destroying the hash table 1160160814Ssimon (and losing the object pointers). So some over-zealous constifications in 1161160814Ssimon LHASH have been relaxed so that lh_insert() does not take (nor store) the 1162160814Ssimon objects as "const" and the lh_doall[_arg] callback wrappers are not 1163160814Ssimon prototyped to have "const" restrictions on the object pointers they are 1164160814Ssimon given (and so aren't required to cast them away any more). 1165160814Ssimon [Geoff Thorpe] 1166160814Ssimon 1167160814Ssimon *) The tmdiff.h API was so ugly and minimal that our own timing utility 1168160814Ssimon (speed) prefers to use its own implementation. The two implementations 1169160814Ssimon haven't been consolidated as yet (volunteers?) but the tmdiff API has had 1170160814Ssimon its object type properly exposed (MS_TM) instead of casting to/from "char 1171160814Ssimon *". This may still change yet if someone realises MS_TM and "ms_time_***" 1172160814Ssimon aren't necessarily the greatest nomenclatures - but this is what was used 1173160814Ssimon internally to the implementation so I've used that for now. 1174160814Ssimon [Geoff Thorpe] 1175160814Ssimon 1176160814Ssimon *) Ensure that deprecated functions do not get compiled when 1177160814Ssimon OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of 1178160814Ssimon the self-tests were still using deprecated key-generation functions so 1179160814Ssimon these have been updated also. 1180160814Ssimon [Geoff Thorpe] 1181160814Ssimon 1182160814Ssimon *) Reorganise PKCS#7 code to separate the digest location functionality 1183160814Ssimon into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest(). 1184160814Ssimon New function PKCS7_set_digest() to set the digest type for PKCS#7 1185160814Ssimon digestedData type. Add additional code to correctly generate the 1186160814Ssimon digestedData type and add support for this type in PKCS7 initialization 1187160814Ssimon functions. 1188160814Ssimon [Steve Henson] 1189160814Ssimon 1190160814Ssimon *) New function PKCS7_set0_type_other() this initializes a PKCS7 1191160814Ssimon structure of type "other". 1192160814Ssimon [Steve Henson] 1193160814Ssimon 1194160814Ssimon *) Fix prime generation loop in crypto/bn/bn_prime.pl by making 1195160814Ssimon sure the loop does correctly stop and breaking ("division by zero") 1196160814Ssimon modulus operations are not performed. The (pre-generated) prime 1197160814Ssimon table crypto/bn/bn_prime.h was already correct, but it could not be 1198160814Ssimon re-generated on some platforms because of the "division by zero" 1199160814Ssimon situation in the script. 1200160814Ssimon [Ralf S. Engelschall] 1201160814Ssimon 1202160814Ssimon *) Update support for ECC-based TLS ciphersuites according to 1203160814Ssimon draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with 1204160814Ssimon SHA-1 now is only used for "small" curves (where the 1205160814Ssimon representation of a field element takes up to 24 bytes); for 1206160814Ssimon larger curves, the field element resulting from ECDH is directly 1207160814Ssimon used as premaster secret. 1208160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 1209160814Ssimon 1210160814Ssimon *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2 1211160814Ssimon curve secp160r1 to the tests. 1212160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 1213160814Ssimon 1214160814Ssimon *) Add the possibility to load symbols globally with DSO. 1215160814Ssimon [G�tz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte] 1216160814Ssimon 1217160814Ssimon *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better 1218160814Ssimon control of the error stack. 1219160814Ssimon [Richard Levitte] 1220160814Ssimon 1221160814Ssimon *) Add support for STORE in ENGINE. 1222160814Ssimon [Richard Levitte] 1223160814Ssimon 1224160814Ssimon *) Add the STORE type. The intention is to provide a common interface 1225160814Ssimon to certificate and key stores, be they simple file-based stores, or 1226160814Ssimon HSM-type store, or LDAP stores, or... 1227160814Ssimon NOTE: The code is currently UNTESTED and isn't really used anywhere. 1228160814Ssimon [Richard Levitte] 1229160814Ssimon 1230160814Ssimon *) Add a generic structure called OPENSSL_ITEM. This can be used to 1231160814Ssimon pass a list of arguments to any function as well as provide a way 1232160814Ssimon for a function to pass data back to the caller. 1233160814Ssimon [Richard Levitte] 1234160814Ssimon 1235160814Ssimon *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup() 1236160814Ssimon works like BUF_strdup() but can be used to duplicate a portion of 1237160814Ssimon a string. The copy gets NUL-terminated. BUF_memdup() duplicates 1238160814Ssimon a memory area. 1239160814Ssimon [Richard Levitte] 1240160814Ssimon 1241160814Ssimon *) Add the function sk_find_ex() which works like sk_find(), but will 1242160814Ssimon return an index to an element even if an exact match couldn't be 1243160814Ssimon found. The index is guaranteed to point at the element where the 1244160814Ssimon searched-for key would be inserted to preserve sorting order. 1245160814Ssimon [Richard Levitte] 1246160814Ssimon 1247160814Ssimon *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but 1248160814Ssimon takes an extra flags argument for optional functionality. Currently, 1249160814Ssimon the following flags are defined: 1250160814Ssimon 1251160814Ssimon OBJ_BSEARCH_VALUE_ON_NOMATCH 1252160814Ssimon This one gets OBJ_bsearch_ex() to return a pointer to the first 1253160814Ssimon element where the comparing function returns a negative or zero 1254160814Ssimon number. 1255160814Ssimon 1256160814Ssimon OBJ_BSEARCH_FIRST_VALUE_ON_MATCH 1257160814Ssimon This one gets OBJ_bsearch_ex() to return a pointer to the first 1258160814Ssimon element where the comparing function returns zero. This is useful 1259160814Ssimon if there are more than one element where the comparing function 1260160814Ssimon returns zero. 1261160814Ssimon [Richard Levitte] 1262160814Ssimon 1263160814Ssimon *) Make it possible to create self-signed certificates with 'openssl ca' 1264160814Ssimon in such a way that the self-signed certificate becomes part of the 1265160814Ssimon CA database and uses the same mechanisms for serial number generation 1266160814Ssimon as all other certificate signing. The new flag '-selfsign' enables 1267160814Ssimon this functionality. Adapt CA.sh and CA.pl.in. 1268160814Ssimon [Richard Levitte] 1269160814Ssimon 1270160814Ssimon *) Add functionality to check the public key of a certificate request 1271160814Ssimon against a given private. This is useful to check that a certificate 1272160814Ssimon request can be signed by that key (self-signing). 1273160814Ssimon [Richard Levitte] 1274160814Ssimon 1275160814Ssimon *) Make it possible to have multiple active certificates with the same 1276160814Ssimon subject in the CA index file. This is done only if the keyword 1277160814Ssimon 'unique_subject' is set to 'no' in the main CA section (default 1278160814Ssimon if 'CA_default') of the configuration file. The value is saved 1279160814Ssimon with the database itself in a separate index attribute file, 1280160814Ssimon named like the index file with '.attr' appended to the name. 1281160814Ssimon [Richard Levitte] 1282160814Ssimon 1283160814Ssimon *) Generate muti valued AVAs using '+' notation in config files for 1284160814Ssimon req and dirName. 1285160814Ssimon [Steve Henson] 1286160814Ssimon 1287160814Ssimon *) Support for nameConstraints certificate extension. 1288160814Ssimon [Steve Henson] 1289160814Ssimon 1290160814Ssimon *) Support for policyConstraints certificate extension. 1291160814Ssimon [Steve Henson] 1292160814Ssimon 1293160814Ssimon *) Support for policyMappings certificate extension. 1294160814Ssimon [Steve Henson] 1295160814Ssimon 1296160814Ssimon *) Make sure the default DSA_METHOD implementation only uses its 1297160814Ssimon dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL, 1298160814Ssimon and change its own handlers to be NULL so as to remove unnecessary 1299160814Ssimon indirection. This lets alternative implementations fallback to the 1300160814Ssimon default implementation more easily. 1301160814Ssimon [Geoff Thorpe] 1302160814Ssimon 1303160814Ssimon *) Support for directoryName in GeneralName related extensions 1304160814Ssimon in config files. 1305160814Ssimon [Steve Henson] 1306160814Ssimon 1307160814Ssimon *) Make it possible to link applications using Makefile.shared. 1308160814Ssimon Make that possible even when linking against static libraries! 1309160814Ssimon [Richard Levitte] 1310160814Ssimon 1311160814Ssimon *) Support for single pass processing for S/MIME signing. This now 1312160814Ssimon means that S/MIME signing can be done from a pipe, in addition 1313160814Ssimon cleartext signing (multipart/signed type) is effectively streaming 1314160814Ssimon and the signed data does not need to be all held in memory. 1315160814Ssimon 1316160814Ssimon This is done with a new flag PKCS7_STREAM. When this flag is set 1317160814Ssimon PKCS7_sign() only initializes the PKCS7 structure and the actual signing 1318160814Ssimon is done after the data is output (and digests calculated) in 1319160814Ssimon SMIME_write_PKCS7(). 1320160814Ssimon [Steve Henson] 1321160814Ssimon 1322160814Ssimon *) Add full support for -rpath/-R, both in shared libraries and 1323160814Ssimon applications, at least on the platforms where it's known how 1324160814Ssimon to do it. 1325160814Ssimon [Richard Levitte] 1326160814Ssimon 1327160814Ssimon *) In crypto/ec/ec_mult.c, implement fast point multiplication with 1328160814Ssimon precomputation, based on wNAF splitting: EC_GROUP_precompute_mult() 1329160814Ssimon will now compute a table of multiples of the generator that 1330160814Ssimon makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul() 1331160814Ssimon faster (notably in the case of a single point multiplication, 1332160814Ssimon scalar * generator). 1333160814Ssimon [Nils Larsch, Bodo Moeller] 1334160814Ssimon 1335160814Ssimon *) IPv6 support for certificate extensions. The various extensions 1336160814Ssimon which use the IP:a.b.c.d can now take IPv6 addresses using the 1337160814Ssimon formats of RFC1884 2.2 . IPv6 addresses are now also displayed 1338160814Ssimon correctly. 1339160814Ssimon [Steve Henson] 1340160814Ssimon 1341160814Ssimon *) Added an ENGINE that implements RSA by performing private key 1342160814Ssimon exponentiations with the GMP library. The conversions to and from 1343160814Ssimon GMP's mpz_t format aren't optimised nor are any montgomery forms 1344160814Ssimon cached, and on x86 it appears OpenSSL's own performance has caught up. 1345160814Ssimon However there are likely to be other architectures where GMP could 1346160814Ssimon provide a boost. This ENGINE is not built in by default, but it can be 1347160814Ssimon specified at Configure time and should be accompanied by the necessary 1348160814Ssimon linker additions, eg; 1349160814Ssimon ./config -DOPENSSL_USE_GMP -lgmp 1350160814Ssimon [Geoff Thorpe] 1351160814Ssimon 1352160814Ssimon *) "openssl engine" will not display ENGINE/DSO load failure errors when 1353160814Ssimon testing availability of engines with "-t" - the old behaviour is 1354160814Ssimon produced by increasing the feature's verbosity with "-tt". 1355160814Ssimon [Geoff Thorpe] 1356160814Ssimon 1357160814Ssimon *) ECDSA routines: under certain error conditions uninitialized BN objects 1358160814Ssimon could be freed. Solution: make sure initialization is performed early 1359160814Ssimon enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de> 1360160814Ssimon via PR#459) 1361160814Ssimon [Lutz Jaenicke] 1362160814Ssimon 1363160814Ssimon *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD 1364160814Ssimon and DH_METHOD (eg. by ENGINE implementations) to override the normal 1365160814Ssimon software implementations. For DSA and DH, parameter generation can 1366160814Ssimon also be overriden by providing the appropriate method callbacks. 1367160814Ssimon [Geoff Thorpe] 1368160814Ssimon 1369160814Ssimon *) Change the "progress" mechanism used in key-generation and 1370160814Ssimon primality testing to functions that take a new BN_GENCB pointer in 1371160814Ssimon place of callback/argument pairs. The new API functions have "_ex" 1372160814Ssimon postfixes and the older functions are reimplemented as wrappers for 1373160814Ssimon the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide 1374160814Ssimon declarations of the old functions to help (graceful) attempts to 1375160814Ssimon migrate to the new functions. Also, the new key-generation API 1376160814Ssimon functions operate on a caller-supplied key-structure and return 1377160814Ssimon success/failure rather than returning a key or NULL - this is to 1378160814Ssimon help make "keygen" another member function of RSA_METHOD etc. 1379160814Ssimon 1380160814Ssimon Example for using the new callback interface: 1381160814Ssimon 1382160814Ssimon int (*my_callback)(int a, int b, BN_GENCB *cb) = ...; 1383160814Ssimon void *my_arg = ...; 1384160814Ssimon BN_GENCB my_cb; 1385160814Ssimon 1386160814Ssimon BN_GENCB_set(&my_cb, my_callback, my_arg); 1387160814Ssimon 1388160814Ssimon return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb); 1389160814Ssimon /* For the meaning of a, b in calls to my_callback(), see the 1390160814Ssimon * documentation of the function that calls the callback. 1391160814Ssimon * cb will point to my_cb; my_arg can be retrieved as cb->arg. 1392160814Ssimon * my_callback should return 1 if it wants BN_is_prime_ex() 1393160814Ssimon * to continue, or 0 to stop. 1394160814Ssimon */ 1395160814Ssimon 1396160814Ssimon [Geoff Thorpe] 1397160814Ssimon 1398160814Ssimon *) Change the ZLIB compression method to be stateful, and make it 1399160814Ssimon available to TLS with the number defined in 1400160814Ssimon draft-ietf-tls-compression-04.txt. 1401160814Ssimon [Richard Levitte] 1402160814Ssimon 1403160814Ssimon *) Add the ASN.1 structures and functions for CertificatePair, which 1404160814Ssimon is defined as follows (according to X.509_4thEditionDraftV6.pdf): 1405160814Ssimon 1406160814Ssimon CertificatePair ::= SEQUENCE { 1407160814Ssimon forward [0] Certificate OPTIONAL, 1408160814Ssimon reverse [1] Certificate OPTIONAL, 1409160814Ssimon -- at least one of the pair shall be present -- } 1410160814Ssimon 1411160814Ssimon Also implement the PEM functions to read and write certificate 1412160814Ssimon pairs, and defined the PEM tag as "CERTIFICATE PAIR". 1413160814Ssimon 1414160814Ssimon This needed to be defined, mostly for the sake of the LDAP 1415160814Ssimon attribute crossCertificatePair, but may prove useful elsewhere as 1416160814Ssimon well. 1417160814Ssimon [Richard Levitte] 1418160814Ssimon 1419160814Ssimon *) Make it possible to inhibit symlinking of shared libraries in 1420160814Ssimon Makefile.shared, for Cygwin's sake. 1421160814Ssimon [Richard Levitte] 1422160814Ssimon 1423160814Ssimon *) Extend the BIGNUM API by creating a function 1424160814Ssimon void BN_set_negative(BIGNUM *a, int neg); 1425160814Ssimon and a macro that behave like 1426160814Ssimon int BN_is_negative(const BIGNUM *a); 1427160814Ssimon 1428160814Ssimon to avoid the need to access 'a->neg' directly in applications. 1429160814Ssimon [Nils Larsch] 1430160814Ssimon 1431160814Ssimon *) Implement fast modular reduction for pseudo-Mersenne primes 1432160814Ssimon used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c). 1433160814Ssimon EC_GROUP_new_curve_GFp() will now automatically use this 1434160814Ssimon if applicable. 1435160814Ssimon [Nils Larsch <nla@trustcenter.de>] 1436160814Ssimon 1437160814Ssimon *) Add new lock type (CRYPTO_LOCK_BN). 1438160814Ssimon [Bodo Moeller] 1439160814Ssimon 1440160814Ssimon *) Change the ENGINE framework to automatically load engines 1441160814Ssimon dynamically from specific directories unless they could be 1442160814Ssimon found to already be built in or loaded. Move all the 1443160814Ssimon current engines except for the cryptodev one to a new 1444160814Ssimon directory engines/. 1445160814Ssimon The engines in engines/ are built as shared libraries if 1446160814Ssimon the "shared" options was given to ./Configure or ./config. 1447160814Ssimon Otherwise, they are inserted in libcrypto.a. 1448160814Ssimon /usr/local/ssl/engines is the default directory for dynamic 1449160814Ssimon engines, but that can be overriden at configure time through 1450160814Ssimon the usual use of --prefix and/or --openssldir, and at run 1451160814Ssimon time with the environment variable OPENSSL_ENGINES. 1452160814Ssimon [Geoff Thorpe and Richard Levitte] 1453160814Ssimon 1454160814Ssimon *) Add Makefile.shared, a helper makefile to build shared 1455160814Ssimon libraries. Addapt Makefile.org. 1456160814Ssimon [Richard Levitte] 1457160814Ssimon 1458160814Ssimon *) Add version info to Win32 DLLs. 1459160814Ssimon [Peter 'Luna' Runestig" <peter@runestig.com>] 1460160814Ssimon 1461160814Ssimon *) Add new 'medium level' PKCS#12 API. Certificates and keys 1462160814Ssimon can be added using this API to created arbitrary PKCS#12 1463160814Ssimon files while avoiding the low level API. 1464160814Ssimon 1465160814Ssimon New options to PKCS12_create(), key or cert can be NULL and 1466160814Ssimon will then be omitted from the output file. The encryption 1467160814Ssimon algorithm NIDs can be set to -1 for no encryption, the mac 1468160814Ssimon iteration count can be set to 0 to omit the mac. 1469160814Ssimon 1470160814Ssimon Enhance pkcs12 utility by making the -nokeys and -nocerts 1471160814Ssimon options work when creating a PKCS#12 file. New option -nomac 1472160814Ssimon to omit the mac, NONE can be set for an encryption algorithm. 1473160814Ssimon New code is modified to use the enhanced PKCS12_create() 1474160814Ssimon instead of the low level API. 1475160814Ssimon [Steve Henson] 1476160814Ssimon 1477160814Ssimon *) Extend ASN1 encoder to support indefinite length constructed 1478160814Ssimon encoding. This can output sequences tags and octet strings in 1479160814Ssimon this form. Modify pk7_asn1.c to support indefinite length 1480160814Ssimon encoding. This is experimental and needs additional code to 1481160814Ssimon be useful, such as an ASN1 bio and some enhanced streaming 1482160814Ssimon PKCS#7 code. 1483160814Ssimon 1484160814Ssimon Extend template encode functionality so that tagging is passed 1485160814Ssimon down to the template encoder. 1486160814Ssimon [Steve Henson] 1487160814Ssimon 1488160814Ssimon *) Let 'openssl req' fail if an argument to '-newkey' is not 1489160814Ssimon recognized instead of using RSA as a default. 1490160814Ssimon [Bodo Moeller] 1491160814Ssimon 1492160814Ssimon *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt. 1493160814Ssimon As these are not official, they are not included in "ALL"; 1494160814Ssimon the "ECCdraft" ciphersuite group alias can be used to select them. 1495160814Ssimon [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)] 1496160814Ssimon 1497160814Ssimon *) Add ECDH engine support. 1498160814Ssimon [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)] 1499160814Ssimon 1500160814Ssimon *) Add ECDH in new directory crypto/ecdh/. 1501160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 1502160814Ssimon 1503160814Ssimon *) Let BN_rand_range() abort with an error after 100 iterations 1504160814Ssimon without success (which indicates a broken PRNG). 1505160814Ssimon [Bodo Moeller] 1506160814Ssimon 1507160814Ssimon *) Change BN_mod_sqrt() so that it verifies that the input value 1508160814Ssimon is really the square of the return value. (Previously, 1509160814Ssimon BN_mod_sqrt would show GIGO behaviour.) 1510160814Ssimon [Bodo Moeller] 1511160814Ssimon 1512160814Ssimon *) Add named elliptic curves over binary fields from X9.62, SECG, 1513160814Ssimon and WAP/WTLS; add OIDs that were still missing. 1514160814Ssimon 1515160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 1516160814Ssimon (Sun Microsystems Laboratories)] 1517160814Ssimon 1518160814Ssimon *) Extend the EC library for elliptic curves over binary fields 1519160814Ssimon (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/). 1520160814Ssimon New EC_METHOD: 1521160814Ssimon 1522160814Ssimon EC_GF2m_simple_method 1523160814Ssimon 1524160814Ssimon New API functions: 1525160814Ssimon 1526160814Ssimon EC_GROUP_new_curve_GF2m 1527160814Ssimon EC_GROUP_set_curve_GF2m 1528160814Ssimon EC_GROUP_get_curve_GF2m 1529160814Ssimon EC_POINT_set_affine_coordinates_GF2m 1530160814Ssimon EC_POINT_get_affine_coordinates_GF2m 1531160814Ssimon EC_POINT_set_compressed_coordinates_GF2m 1532160814Ssimon 1533160814Ssimon Point compression for binary fields is disabled by default for 1534160814Ssimon patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to 1535160814Ssimon enable it). 1536160814Ssimon 1537160814Ssimon As binary polynomials are represented as BIGNUMs, various members 1538160814Ssimon of the EC_GROUP and EC_POINT data structures can be shared 1539160814Ssimon between the implementations for prime fields and binary fields; 1540160814Ssimon the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m) 1541160814Ssimon are essentially identical to their ..._GFp counterparts. 1542160814Ssimon (For simplicity, the '..._GFp' prefix has been dropped from 1543160814Ssimon various internal method names.) 1544160814Ssimon 1545160814Ssimon An internal 'field_div' method (similar to 'field_mul' and 1546160814Ssimon 'field_sqr') has been added; this is used only for binary fields. 1547160814Ssimon 1548160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 1549160814Ssimon (Sun Microsystems Laboratories)] 1550160814Ssimon 1551160814Ssimon *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult() 1552160814Ssimon through methods ('mul', 'precompute_mult'). 1553160814Ssimon 1554160814Ssimon The generic implementations (now internally called 'ec_wNAF_mul' 1555160814Ssimon and 'ec_wNAF_precomputed_mult') remain the default if these 1556160814Ssimon methods are undefined. 1557160814Ssimon 1558160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 1559160814Ssimon (Sun Microsystems Laboratories)] 1560160814Ssimon 1561160814Ssimon *) New function EC_GROUP_get_degree, which is defined through 1562160814Ssimon EC_METHOD. For curves over prime fields, this returns the bit 1563160814Ssimon length of the modulus. 1564160814Ssimon 1565160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 1566160814Ssimon (Sun Microsystems Laboratories)] 1567160814Ssimon 1568160814Ssimon *) New functions EC_GROUP_dup, EC_POINT_dup. 1569160814Ssimon (These simply call ..._new and ..._copy). 1570160814Ssimon 1571160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 1572160814Ssimon (Sun Microsystems Laboratories)] 1573160814Ssimon 1574160814Ssimon *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c. 1575160814Ssimon Polynomials are represented as BIGNUMs (where the sign bit is not 1576160814Ssimon used) in the following functions [macros]: 1577160814Ssimon 1578160814Ssimon BN_GF2m_add 1579160814Ssimon BN_GF2m_sub [= BN_GF2m_add] 1580160814Ssimon BN_GF2m_mod [wrapper for BN_GF2m_mod_arr] 1581160814Ssimon BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr] 1582160814Ssimon BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr] 1583160814Ssimon BN_GF2m_mod_inv 1584160814Ssimon BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr] 1585160814Ssimon BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr] 1586160814Ssimon BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr] 1587160814Ssimon BN_GF2m_cmp [= BN_ucmp] 1588160814Ssimon 1589160814Ssimon (Note that only the 'mod' functions are actually for fields GF(2^m). 1590160814Ssimon BN_GF2m_add() is misnomer, but this is for the sake of consistency.) 1591160814Ssimon 1592160814Ssimon For some functions, an the irreducible polynomial defining a 1593160814Ssimon field can be given as an 'unsigned int[]' with strictly 1594160814Ssimon decreasing elements giving the indices of those bits that are set; 1595160814Ssimon i.e., p[] represents the polynomial 1596160814Ssimon f(t) = t^p[0] + t^p[1] + ... + t^p[k] 1597160814Ssimon where 1598160814Ssimon p[0] > p[1] > ... > p[k] = 0. 1599160814Ssimon This applies to the following functions: 1600160814Ssimon 1601160814Ssimon BN_GF2m_mod_arr 1602160814Ssimon BN_GF2m_mod_mul_arr 1603160814Ssimon BN_GF2m_mod_sqr_arr 1604160814Ssimon BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv] 1605160814Ssimon BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div] 1606160814Ssimon BN_GF2m_mod_exp_arr 1607160814Ssimon BN_GF2m_mod_sqrt_arr 1608160814Ssimon BN_GF2m_mod_solve_quad_arr 1609160814Ssimon BN_GF2m_poly2arr 1610160814Ssimon BN_GF2m_arr2poly 1611160814Ssimon 1612160814Ssimon Conversion can be performed by the following functions: 1613160814Ssimon 1614160814Ssimon BN_GF2m_poly2arr 1615160814Ssimon BN_GF2m_arr2poly 1616160814Ssimon 1617160814Ssimon bntest.c has additional tests for binary polynomial arithmetic. 1618160814Ssimon 1619160814Ssimon Two implementations for BN_GF2m_mod_div() are available. 1620160814Ssimon The default algorithm simply uses BN_GF2m_mod_inv() and 1621160814Ssimon BN_GF2m_mod_mul(). The alternative algorithm is compiled in only 1622160814Ssimon if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the 1623160814Ssimon copyright notice in crypto/bn/bn_gf2m.c before enabling it). 1624160814Ssimon 1625160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 1626160814Ssimon (Sun Microsystems Laboratories)] 1627160814Ssimon 1628160814Ssimon *) Add new error code 'ERR_R_DISABLED' that can be used when some 1629160814Ssimon functionality is disabled at compile-time. 1630160814Ssimon [Douglas Stebila <douglas.stebila@sun.com>] 1631160814Ssimon 1632160814Ssimon *) Change default behaviour of 'openssl asn1parse' so that more 1633160814Ssimon information is visible when viewing, e.g., a certificate: 1634160814Ssimon 1635160814Ssimon Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump' 1636160814Ssimon mode the content of non-printable OCTET STRINGs is output in a 1637160814Ssimon style similar to INTEGERs, but with '[HEX DUMP]' prepended to 1638160814Ssimon avoid the appearance of a printable string. 1639160814Ssimon [Nils Larsch <nla@trustcenter.de>] 1640160814Ssimon 1641160814Ssimon *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access 1642160814Ssimon functions 1643160814Ssimon EC_GROUP_set_asn1_flag() 1644160814Ssimon EC_GROUP_get_asn1_flag() 1645160814Ssimon EC_GROUP_set_point_conversion_form() 1646160814Ssimon EC_GROUP_get_point_conversion_form() 1647160814Ssimon These control ASN1 encoding details: 1648160814Ssimon - Curves (i.e., groups) are encoded explicitly unless asn1_flag 1649160814Ssimon has been set to OPENSSL_EC_NAMED_CURVE. 1650160814Ssimon - Points are encoded in uncompressed form by default; options for 1651160814Ssimon asn1_for are as for point2oct, namely 1652160814Ssimon POINT_CONVERSION_COMPRESSED 1653160814Ssimon POINT_CONVERSION_UNCOMPRESSED 1654160814Ssimon POINT_CONVERSION_HYBRID 1655160814Ssimon 1656160814Ssimon Also add 'seed' and 'seed_len' members to EC_GROUP with access 1657160814Ssimon functions 1658160814Ssimon EC_GROUP_set_seed() 1659160814Ssimon EC_GROUP_get0_seed() 1660160814Ssimon EC_GROUP_get_seed_len() 1661160814Ssimon This is used only for ASN1 purposes (so far). 1662160814Ssimon [Nils Larsch <nla@trustcenter.de>] 1663160814Ssimon 1664160814Ssimon *) Add 'field_type' member to EC_METHOD, which holds the NID 1665160814Ssimon of the appropriate field type OID. The new function 1666160814Ssimon EC_METHOD_get_field_type() returns this value. 1667160814Ssimon [Nils Larsch <nla@trustcenter.de>] 1668160814Ssimon 1669160814Ssimon *) Add functions 1670160814Ssimon EC_POINT_point2bn() 1671160814Ssimon EC_POINT_bn2point() 1672160814Ssimon EC_POINT_point2hex() 1673160814Ssimon EC_POINT_hex2point() 1674160814Ssimon providing useful interfaces to EC_POINT_point2oct() and 1675160814Ssimon EC_POINT_oct2point(). 1676160814Ssimon [Nils Larsch <nla@trustcenter.de>] 1677160814Ssimon 1678160814Ssimon *) Change internals of the EC library so that the functions 1679160814Ssimon EC_GROUP_set_generator() 1680160814Ssimon EC_GROUP_get_generator() 1681160814Ssimon EC_GROUP_get_order() 1682160814Ssimon EC_GROUP_get_cofactor() 1683160814Ssimon are implemented directly in crypto/ec/ec_lib.c and not dispatched 1684160814Ssimon to methods, which would lead to unnecessary code duplication when 1685160814Ssimon adding different types of curves. 1686160814Ssimon [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller] 1687160814Ssimon 1688160814Ssimon *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM 1689160814Ssimon arithmetic, and such that modified wNAFs are generated 1690160814Ssimon (which avoid length expansion in many cases). 1691160814Ssimon [Bodo Moeller] 1692160814Ssimon 1693160814Ssimon *) Add a function EC_GROUP_check_discriminant() (defined via 1694160814Ssimon EC_METHOD) that verifies that the curve discriminant is non-zero. 1695160814Ssimon 1696160814Ssimon Add a function EC_GROUP_check() that makes some sanity tests 1697160814Ssimon on a EC_GROUP, its generator and order. This includes 1698160814Ssimon EC_GROUP_check_discriminant(). 1699160814Ssimon [Nils Larsch <nla@trustcenter.de>] 1700160814Ssimon 1701160814Ssimon *) Add ECDSA in new directory crypto/ecdsa/. 1702160814Ssimon 1703160814Ssimon Add applications 'openssl ecparam' and 'openssl ecdsa' 1704160814Ssimon (these are based on 'openssl dsaparam' and 'openssl dsa'). 1705160814Ssimon 1706160814Ssimon ECDSA support is also included in various other files across the 1707160814Ssimon library. Most notably, 1708160814Ssimon - 'openssl req' now has a '-newkey ecdsa:file' option; 1709160814Ssimon - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA; 1710160814Ssimon - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and 1711160814Ssimon d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make 1712160814Ssimon them suitable for ECDSA where domain parameters must be 1713160814Ssimon extracted before the specific public key; 1714160814Ssimon - ECDSA engine support has been added. 1715160814Ssimon [Nils Larsch <nla@trustcenter.de>] 1716160814Ssimon 1717160814Ssimon *) Include some named elliptic curves, and add OIDs from X9.62, 1718160814Ssimon SECG, and WAP/WTLS. Each curve can be obtained from the new 1719160814Ssimon function 1720160814Ssimon EC_GROUP_new_by_curve_name(), 1721160814Ssimon and the list of available named curves can be obtained with 1722160814Ssimon EC_get_builtin_curves(). 1723160814Ssimon Also add a 'curve_name' member to EC_GROUP objects, which can be 1724160814Ssimon accessed via 1725160814Ssimon EC_GROUP_set_curve_name() 1726160814Ssimon EC_GROUP_get_curve_name() 1727160814Ssimon [Nils Larsch <larsch@trustcenter.de, Bodo Moeller] 1728160814Ssimon 1729160814Ssimon *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 1730160814Ssimon was actually never needed) and in BN_mul(). The removal in BN_mul() 1731160814Ssimon required a small change in bn_mul_part_recursive() and the addition 1732160814Ssimon of the functions bn_cmp_part_words(), bn_sub_part_words() and 1733160814Ssimon bn_add_part_words(), which do the same thing as bn_cmp_words(), 1734160814Ssimon bn_sub_words() and bn_add_words() except they take arrays with 1735160814Ssimon differing sizes. 1736160814Ssimon [Richard Levitte] 1737160814Ssimon 1738194206Ssimon Changes between 0.9.7l and 0.9.7m [23 Feb 2007] 1739194206Ssimon 1740167612Ssimon *) Cleanse PEM buffers before freeing them since they may contain 1741167612Ssimon sensitive data. 1742167612Ssimon [Benjamin Bennett <ben@psc.edu>] 1743167612Ssimon 1744167612Ssimon *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 1745167612Ssimon a ciphersuite string such as "DEFAULT:RSA" cannot enable 1746167612Ssimon authentication-only ciphersuites. 1747167612Ssimon [Bodo Moeller] 1748167612Ssimon 1749167612Ssimon *) Since AES128 and AES256 share a single mask bit in the logic of 1750167612Ssimon ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 1751167612Ssimon kludge to work properly if AES128 is available and AES256 isn't. 1752167612Ssimon [Victor Duchovni] 1753167612Ssimon 1754194206Ssimon *) Expand security boundary to match 1.1.1 module. 1755194206Ssimon [Steve Henson] 1756194206Ssimon 1757194206Ssimon *) Remove redundant features: hash file source, editing of test vectors 1758194206Ssimon modify fipsld to use external fips_premain.c signature. 1759194206Ssimon [Steve Henson] 1760194206Ssimon 1761194206Ssimon *) New perl script mkfipsscr.pl to create shell scripts or batch files to 1762194206Ssimon run algorithm test programs. 1763194206Ssimon [Steve Henson] 1764194206Ssimon 1765194206Ssimon *) Make algorithm test programs more tolerant of whitespace. 1766194206Ssimon [Steve Henson] 1767194206Ssimon 1768167612Ssimon *) Have SSL/TLS server implementation tolerate "mismatched" record 1769167612Ssimon protocol version while receiving ClientHello even if the 1770167612Ssimon ClientHello is fragmented. (The server can't insist on the 1771167612Ssimon particular protocol version it has chosen before the ServerHello 1772167612Ssimon message has informed the client about his choice.) 1773167612Ssimon [Bodo Moeller] 1774167612Ssimon 1775167612Ssimon *) Load error codes if they are not already present instead of using a 1776167612Ssimon static variable. This allows them to be cleanly unloaded and reloaded. 1777167612Ssimon [Steve Henson] 1778167612Ssimon 1779167612Ssimon Changes between 0.9.7k and 0.9.7l [28 Sep 2006] 1780167612Ssimon 1781167612Ssimon *) Introduce limits to prevent malicious keys being able to 1782167612Ssimon cause a denial of service. (CVE-2006-2940) 1783167612Ssimon [Steve Henson, Bodo Moeller] 1784167612Ssimon 1785167612Ssimon *) Fix ASN.1 parsing of certain invalid structures that can result 1786167612Ssimon in a denial of service. (CVE-2006-2937) [Steve Henson] 1787167612Ssimon 1788167612Ssimon *) Fix buffer overflow in SSL_get_shared_ciphers() function. 1789167612Ssimon (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 1790167612Ssimon 1791167612Ssimon *) Fix SSL client code which could crash if connecting to a 1792167612Ssimon malicious SSLv2 server. (CVE-2006-4343) 1793167612Ssimon [Tavis Ormandy and Will Drewry, Google Security Team] 1794167612Ssimon 1795162911Ssimon *) Change ciphersuite string processing so that an explicit 1796162911Ssimon ciphersuite selects this one ciphersuite (so that "AES256-SHA" 1797162911Ssimon will no longer include "AES128-SHA"), and any other similar 1798162911Ssimon ciphersuite (same bitmap) from *other* protocol versions (so that 1799162911Ssimon "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the 1800162911Ssimon SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining 1801162911Ssimon changes from 0.9.8b and 0.9.8d. 1802162911Ssimon [Bodo Moeller] 1803162911Ssimon 1804162911Ssimon Changes between 0.9.7j and 0.9.7k [05 Sep 2006] 1805162911Ssimon 1806162911Ssimon *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 1807162911Ssimon (CVE-2006-4339) [Ben Laurie and Google Security Team] 1808162911Ssimon 1809162911Ssimon *) Change the Unix randomness entropy gathering to use poll() when 1810162911Ssimon possible instead of select(), since the latter has some 1811162911Ssimon undesirable limitations. 1812162911Ssimon [Darryl Miles via Richard Levitte and Bodo Moeller] 1813162911Ssimon 1814162911Ssimon *) Disable rogue ciphersuites: 1815162911Ssimon 1816162911Ssimon - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 1817162911Ssimon - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 1818162911Ssimon - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 1819162911Ssimon 1820162911Ssimon The latter two were purportedly from 1821162911Ssimon draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 1822162911Ssimon appear there. 1823162911Ssimon 1824162911Ssimon Also deactive the remaining ciphersuites from 1825162911Ssimon draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 1826162911Ssimon unofficial, and the ID has long expired. 1827162911Ssimon [Bodo Moeller] 1828162911Ssimon 1829162911Ssimon *) Fix RSA blinding Heisenbug (problems sometimes occured on 1830162911Ssimon dual-core machines) and other potential thread-safety issues. 1831162911Ssimon [Bodo Moeller] 1832162911Ssimon 1833162911Ssimon Changes between 0.9.7i and 0.9.7j [04 May 2006] 1834162911Ssimon 1835162911Ssimon *) Adapt fipsld and the build system to link against the validated FIPS 1836162911Ssimon module in FIPS mode. 1837162911Ssimon [Steve Henson] 1838162911Ssimon 1839162911Ssimon *) Fixes for VC++ 2005 build under Windows. 1840162911Ssimon [Steve Henson] 1841162911Ssimon 1842162911Ssimon *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make 1843162911Ssimon from a Windows bash shell such as MSYS. It is autodetected from the 1844162911Ssimon "config" script when run from a VC++ environment. Modify standard VC++ 1845162911Ssimon build to use fipscanister.o from the GNU make build. 1846162911Ssimon [Steve Henson] 1847162911Ssimon 1848160814Ssimon Changes between 0.9.7h and 0.9.7i [14 Oct 2005] 1849160814Ssimon 1850160814Ssimon *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS. 1851160814Ssimon The value now differs depending on if you build for FIPS or not. 1852160814Ssimon BEWARE! A program linked with a shared FIPSed libcrypto can't be 1853160814Ssimon safely run with a non-FIPSed libcrypto, as it may crash because of 1854160814Ssimon the difference induced by this change. 1855160814Ssimon [Andy Polyakov] 1856160814Ssimon 1857160814Ssimon Changes between 0.9.7g and 0.9.7h [11 Oct 2005] 1858160814Ssimon 1859160814Ssimon *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 1860160814Ssimon (part of SSL_OP_ALL). This option used to disable the 1861160814Ssimon countermeasure against man-in-the-middle protocol-version 1862160814Ssimon rollback in the SSL 2.0 server implementation, which is a bad 1863160814Ssimon idea. (CVE-2005-2969) 1864160814Ssimon 1865160814Ssimon [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 1866160814Ssimon for Information Security, National Institute of Advanced Industrial 1867160814Ssimon Science and Technology [AIST], Japan)] 1868160814Ssimon 1869160814Ssimon *) Minimal support for X9.31 signatures and PSS padding modes. This is 1870160814Ssimon mainly for FIPS compliance and not fully integrated at this stage. 1871160814Ssimon [Steve Henson] 1872160814Ssimon 1873160814Ssimon *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform 1874160814Ssimon the exponentiation using a fixed-length exponent. (Otherwise, 1875160814Ssimon the information leaked through timing could expose the secret key 1876160814Ssimon after many signatures; cf. Bleichenbacher's attack on DSA with 1877160814Ssimon biased k.) 1878160814Ssimon [Bodo Moeller] 1879160814Ssimon 1880160814Ssimon *) Make a new fixed-window mod_exp implementation the default for 1881160814Ssimon RSA, DSA, and DH private-key operations so that the sequence of 1882160814Ssimon squares and multiplies and the memory access pattern are 1883160814Ssimon independent of the particular secret key. This will mitigate 1884160814Ssimon cache-timing and potential related attacks. 1885160814Ssimon 1886160814Ssimon BN_mod_exp_mont_consttime() is the new exponentiation implementation, 1887160814Ssimon and this is automatically used by BN_mod_exp_mont() if the new flag 1888160814Ssimon BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH 1889160814Ssimon will use this BN flag for private exponents unless the flag 1890160814Ssimon RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or 1891160814Ssimon DH_FLAG_NO_EXP_CONSTTIME, respectively, is set. 1892160814Ssimon 1893160814Ssimon [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller] 1894160814Ssimon 1895160814Ssimon *) Change the client implementation for SSLv23_method() and 1896160814Ssimon SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0 1897160814Ssimon Client Hello message format if the SSL_OP_NO_SSLv2 option is set. 1898160814Ssimon (Previously, the SSL 2.0 backwards compatible Client Hello 1899160814Ssimon message format would be used even with SSL_OP_NO_SSLv2.) 1900160814Ssimon [Bodo Moeller] 1901160814Ssimon 1902160814Ssimon *) Add support for smime-type MIME parameter in S/MIME messages which some 1903160814Ssimon clients need. 1904160814Ssimon [Steve Henson] 1905160814Ssimon 1906160814Ssimon *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in 1907160814Ssimon a threadsafe manner. Modify rsa code to use new function and add calls 1908160814Ssimon to dsa and dh code (which had race conditions before). 1909160814Ssimon [Steve Henson] 1910160814Ssimon 1911160814Ssimon *) Include the fixed error library code in the C error file definitions 1912160814Ssimon instead of fixing them up at runtime. This keeps the error code 1913160814Ssimon structures constant. 1914160814Ssimon [Steve Henson] 1915160814Ssimon 1916160814Ssimon Changes between 0.9.7f and 0.9.7g [11 Apr 2005] 1917160814Ssimon 1918160814Ssimon [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after 1919160814Ssimon OpenSSL 0.9.8.] 1920160814Ssimon 1921160814Ssimon *) Fixes for newer kerberos headers. NB: the casts are needed because 1922160814Ssimon the 'length' field is signed on one version and unsigned on another 1923160814Ssimon with no (?) obvious way to tell the difference, without these VC++ 1924160814Ssimon complains. Also the "definition" of FAR (blank) is no longer included 1925160814Ssimon nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up 1926160814Ssimon some needed definitions. 1927160814Ssimon [Steve Henson] 1928160814Ssimon 1929160814Ssimon *) Undo Cygwin change. 1930160814Ssimon [Ulf M�ller] 1931160814Ssimon 1932160814Ssimon *) Added support for proxy certificates according to RFC 3820. 1933160814Ssimon Because they may be a security thread to unaware applications, 1934160814Ssimon they must be explicitely allowed in run-time. See 1935160814Ssimon docs/HOWTO/proxy_certificates.txt for further information. 1936160814Ssimon [Richard Levitte] 1937160814Ssimon 1938160814Ssimon Changes between 0.9.7e and 0.9.7f [22 Mar 2005] 1939160814Ssimon 1940160814Ssimon *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating 1941160814Ssimon server and client random values. Previously 1942160814Ssimon (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in 1943160814Ssimon less random data when sizeof(time_t) > 4 (some 64 bit platforms). 1944160814Ssimon 1945160814Ssimon This change has negligible security impact because: 1946160814Ssimon 1947160814Ssimon 1. Server and client random values still have 24 bytes of pseudo random 1948160814Ssimon data. 1949160814Ssimon 1950160814Ssimon 2. Server and client random values are sent in the clear in the initial 1951160814Ssimon handshake. 1952160814Ssimon 1953160814Ssimon 3. The master secret is derived using the premaster secret (48 bytes in 1954160814Ssimon size for static RSA ciphersuites) as well as client server and random 1955160814Ssimon values. 1956160814Ssimon 1957160814Ssimon The OpenSSL team would like to thank the UK NISCC for bringing this issue 1958160814Ssimon to our attention. 1959160814Ssimon 1960160814Ssimon [Stephen Henson, reported by UK NISCC] 1961160814Ssimon 1962160814Ssimon *) Use Windows randomness collection on Cygwin. 1963160814Ssimon [Ulf M�ller] 1964160814Ssimon 1965160814Ssimon *) Fix hang in EGD/PRNGD query when communication socket is closed 1966160814Ssimon prematurely by EGD/PRNGD. 1967160814Ssimon [Darren Tucker <dtucker@zip.com.au> via Lutz J�nicke, resolves #1014] 1968160814Ssimon 1969160814Ssimon *) Prompt for pass phrases when appropriate for PKCS12 input format. 1970160814Ssimon [Steve Henson] 1971160814Ssimon 1972160814Ssimon *) Back-port of selected performance improvements from development 1973160814Ssimon branch, as well as improved support for PowerPC platforms. 1974160814Ssimon [Andy Polyakov] 1975160814Ssimon 1976160814Ssimon *) Add lots of checks for memory allocation failure, error codes to indicate 1977160814Ssimon failure and freeing up memory if a failure occurs. 1978160814Ssimon [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson] 1979160814Ssimon 1980160814Ssimon *) Add new -passin argument to dgst. 1981160814Ssimon [Steve Henson] 1982160814Ssimon 1983160814Ssimon *) Perform some character comparisons of different types in X509_NAME_cmp: 1984160814Ssimon this is needed for some certificates that reencode DNs into UTF8Strings 1985160814Ssimon (in violation of RFC3280) and can't or wont issue name rollover 1986160814Ssimon certificates. 1987160814Ssimon [Steve Henson] 1988160814Ssimon 1989160814Ssimon *) Make an explicit check during certificate validation to see that 1990160814Ssimon the CA setting in each certificate on the chain is correct. As a 1991160814Ssimon side effect always do the following basic checks on extensions, 1992160814Ssimon not just when there's an associated purpose to the check: 1993160814Ssimon 1994160814Ssimon - if there is an unhandled critical extension (unless the user 1995160814Ssimon has chosen to ignore this fault) 1996160814Ssimon - if the path length has been exceeded (if one is set at all) 1997160814Ssimon - that certain extensions fit the associated purpose (if one has 1998160814Ssimon been given) 1999160814Ssimon [Richard Levitte] 2000160814Ssimon 2001142425Snectar Changes between 0.9.7d and 0.9.7e [25 Oct 2004] 2002142425Snectar 2003142425Snectar *) Avoid a race condition when CRLs are checked in a multi threaded 2004142425Snectar environment. This would happen due to the reordering of the revoked 2005142425Snectar entries during signature checking and serial number lookup. Now the 2006142425Snectar encoding is cached and the serial number sort performed under a lock. 2007142425Snectar Add new STACK function sk_is_sorted(). 2008142425Snectar [Steve Henson] 2009142425Snectar 2010142425Snectar *) Add Delta CRL to the extension code. 2011142425Snectar [Steve Henson] 2012142425Snectar 2013142425Snectar *) Various fixes to s3_pkt.c so alerts are sent properly. 2014142425Snectar [David Holmes <d.holmes@f5.com>] 2015142425Snectar 2016142425Snectar *) Reduce the chances of duplicate issuer name and serial numbers (in 2017142425Snectar violation of RFC3280) using the OpenSSL certificate creation utilities. 2018142425Snectar This is done by creating a random 64 bit value for the initial serial 2019142425Snectar number when a serial number file is created or when a self signed 2020142425Snectar certificate is created using 'openssl req -x509'. The initial serial 2021142425Snectar number file is created using 'openssl x509 -next_serial' in CA.pl 2022142425Snectar rather than being initialized to 1. 2023142425Snectar [Steve Henson] 2024142425Snectar 2025127128Snectar Changes between 0.9.7c and 0.9.7d [17 Mar 2004] 2026127128Snectar 2027127128Snectar *) Fix null-pointer assignment in do_change_cipher_spec() revealed 2028160814Ssimon by using the Codenomicon TLS Test Tool (CVE-2004-0079) 2029127128Snectar [Joe Orton, Steve Henson] 2030127128Snectar 2031127128Snectar *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites 2032160814Ssimon (CVE-2004-0112) 2033127128Snectar [Joe Orton, Steve Henson] 2034127128Snectar 2035127128Snectar *) Make it possible to have multiple active certificates with the same 2036127128Snectar subject in the CA index file. This is done only if the keyword 2037127128Snectar 'unique_subject' is set to 'no' in the main CA section (default 2038127128Snectar if 'CA_default') of the configuration file. The value is saved 2039127128Snectar with the database itself in a separate index attribute file, 2040127128Snectar named like the index file with '.attr' appended to the name. 2041127128Snectar [Richard Levitte] 2042127128Snectar 2043127128Snectar *) X509 verify fixes. Disable broken certificate workarounds when 2044127128Snectar X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if 2045127128Snectar keyUsage extension present. Don't accept CRLs with unhandled critical 2046127128Snectar extensions: since verify currently doesn't process CRL extensions this 2047127128Snectar rejects a CRL with *any* critical extensions. Add new verify error codes 2048127128Snectar for these cases. 2049127128Snectar [Steve Henson] 2050127128Snectar 2051127128Snectar *) When creating an OCSP nonce use an OCTET STRING inside the extnValue. 2052127128Snectar A clarification of RFC2560 will require the use of OCTET STRINGs and 2053127128Snectar some implementations cannot handle the current raw format. Since OpenSSL 2054127128Snectar copies and compares OCSP nonces as opaque blobs without any attempt at 2055127128Snectar parsing them this should not create any compatibility issues. 2056127128Snectar [Steve Henson] 2057127128Snectar 2058127128Snectar *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when 2059127128Snectar calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without 2060127128Snectar this HMAC (and other) operations are several times slower than OpenSSL 2061127128Snectar < 0.9.7. 2062127128Snectar [Steve Henson] 2063127128Snectar 2064127128Snectar *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex(). 2065127128Snectar [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 2066127128Snectar 2067127128Snectar *) Use the correct content when signing type "other". 2068127128Snectar [Steve Henson] 2069127128Snectar 2070120631Snectar Changes between 0.9.7b and 0.9.7c [30 Sep 2003] 2071120631Snectar 2072120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 2073120631Snectar 2074120631Snectar Stop out of bounds reads in the ASN1 code when presented with 2075160814Ssimon invalid tags (CVE-2003-0543 and CVE-2003-0544). 2076120631Snectar 2077160814Ssimon Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545). 2078120631Snectar 2079120631Snectar If verify callback ignores invalid public key errors don't try to check 2080120631Snectar certificate signature with the NULL public key. 2081120631Snectar 2082120631Snectar [Steve Henson] 2083120631Snectar 2084120631Snectar *) New -ignore_err option in ocsp application to stop the server 2085120631Snectar exiting on the first error in a request. 2086120631Snectar [Steve Henson] 2087120631Snectar 2088120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 2089120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 2090120631Snectar specifications. 2091120631Snectar [Steve Henson] 2092120631Snectar 2093120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 2094120631Snectar extra data after the compression methods not only for TLS 1.0 2095120631Snectar but also for SSL 3.0 (as required by the specification). 2096120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 2097120631Snectar 2098120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 2099120631Snectar when it's 512 *bits* long, not 512 bytes. 2100120631Snectar [Richard Levitte] 2101120631Snectar 2102120631Snectar *) Change AES_cbc_encrypt() so it outputs exact multiple of 2103120631Snectar blocks during encryption. 2104120631Snectar [Richard Levitte] 2105120631Snectar 2106120631Snectar *) Various fixes to base64 BIO and non blocking I/O. On write 2107120631Snectar flushes were not handled properly if the BIO retried. On read 2108120631Snectar data was not being buffered properly and had various logic bugs. 2109120631Snectar This also affects blocking I/O when the data being decoded is a 2110120631Snectar certain size. 2111120631Snectar [Steve Henson] 2112120631Snectar 2113120631Snectar *) Various S/MIME bugfixes and compatibility changes: 2114120631Snectar output correct application/pkcs7 MIME type if 2115120631Snectar PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures. 2116120631Snectar Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening 2117120631Snectar of files as .eml work). Correctly handle very long lines in MIME 2118120631Snectar parser. 2119120631Snectar [Steve Henson] 2120120631Snectar 2121120631Snectar Changes between 0.9.7a and 0.9.7b [10 Apr 2003] 2122120631Snectar 2123120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 2124120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 2125120631Snectar a protocol version number mismatch like a decryption error 2126120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 2127120631Snectar [Bodo Moeller] 2128120631Snectar 2129120631Snectar *) Turn on RSA blinding by default in the default implementation 2130120631Snectar to avoid a timing attack. Applications that don't want it can call 2131120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 2132120631Snectar They would be ill-advised to do so in most cases. 2133120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 2134120631Snectar 2135120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 2136120631Snectar seeded (in this case, the secret RSA exponent is abused as 2137120631Snectar an unpredictable seed -- if it is not unpredictable, there 2138120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 2139120631Snectar by remembering the creator's thread ID in rsa->blinding and 2140120631Snectar having all other threads use local one-time blinding factors 2141120631Snectar (this requires more computation than sharing rsa->blinding, but 2142120631Snectar avoids excessive locking; and if an RSA object is not shared 2143120631Snectar between threads, blinding will still be very fast). 2144120631Snectar [Bodo Moeller] 2145120631Snectar 2146120631Snectar *) Fixed a typo bug that would cause ENGINE_set_default() to set an 2147120631Snectar ENGINE as defaults for all supported algorithms irrespective of 2148120631Snectar the 'flags' parameter. 'flags' is now honoured, so applications 2149120631Snectar should make sure they are passing it correctly. 2150120631Snectar [Geoff Thorpe] 2151120631Snectar 2152120631Snectar *) Target "mingw" now allows native Windows code to be generated in 2153120631Snectar the Cygwin environment as well as with the MinGW compiler. 2154120631Snectar [Ulf Moeller] 2155120631Snectar 2156111147Snectar Changes between 0.9.7 and 0.9.7a [19 Feb 2003] 2157111147Snectar 2158111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 2159111147Snectar via timing by performing a MAC computation even if incorrrect 2160111147Snectar block cipher padding has been found. This is a countermeasure 2161111147Snectar against active attacks where the attacker has to distinguish 2162160814Ssimon between bad padding and a MAC verification error. (CVE-2003-0078) 2163111147Snectar 2164111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 2165111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 2166111147Snectar Martin Vuagnoux (EPFL, Ilion)] 2167111147Snectar 2168111147Snectar *) Make the no-err option work as intended. The intention with no-err 2169111147Snectar is not to have the whole error stack handling routines removed from 2170111147Snectar libcrypto, it's only intended to remove all the function name and 2171111147Snectar reason texts, thereby removing some of the footprint that may not 2172111147Snectar be interesting if those errors aren't displayed anyway. 2173111147Snectar 2174111147Snectar NOTE: it's still possible for any application or module to have it's 2175111147Snectar own set of error texts inserted. The routines are there, just not 2176111147Snectar used by default when no-err is given. 2177111147Snectar [Richard Levitte] 2178111147Snectar 2179111147Snectar *) Add support for FreeBSD on IA64. 2180111147Snectar [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454] 2181111147Snectar 2182111147Snectar *) Adjust DES_cbc_cksum() so it returns the same value as the MIT 2183111147Snectar Kerberos function mit_des_cbc_cksum(). Before this change, 2184111147Snectar the value returned by DES_cbc_cksum() was like the one from 2185111147Snectar mit_des_cbc_cksum(), except the bytes were swapped. 2186111147Snectar [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte] 2187111147Snectar 2188111147Snectar *) Allow an application to disable the automatic SSL chain building. 2189111147Snectar Before this a rather primitive chain build was always performed in 2190111147Snectar ssl3_output_cert_chain(): an application had no way to send the 2191111147Snectar correct chain if the automatic operation produced an incorrect result. 2192111147Snectar 2193111147Snectar Now the chain builder is disabled if either: 2194111147Snectar 2195111147Snectar 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert(). 2196111147Snectar 2197111147Snectar 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set. 2198111147Snectar 2199111147Snectar The reasoning behind this is that an application would not want the 2200111147Snectar auto chain building to take place if extra chain certificates are 2201111147Snectar present and it might also want a means of sending no additional 2202111147Snectar certificates (for example the chain has two certificates and the 2203111147Snectar root is omitted). 2204111147Snectar [Steve Henson] 2205111147Snectar 2206111147Snectar *) Add the possibility to build without the ENGINE framework. 2207111147Snectar [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 2208111147Snectar 2209111147Snectar *) Under Win32 gmtime() can return NULL: check return value in 2210111147Snectar OPENSSL_gmtime(). Add error code for case where gmtime() fails. 2211111147Snectar [Steve Henson] 2212111147Snectar 2213111147Snectar *) DSA routines: under certain error conditions uninitialized BN objects 2214111147Snectar could be freed. Solution: make sure initialization is performed early 2215111147Snectar enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>, 2216111147Snectar Nils Larsch <nla@trustcenter.de> via PR#459) 2217111147Snectar [Lutz Jaenicke] 2218111147Snectar 2219111147Snectar *) Another fix for SSLv2 session ID handling: the session ID was incorrectly 2220111147Snectar checked on reconnect on the client side, therefore session resumption 2221111147Snectar could still fail with a "ssl session id is different" error. This 2222111147Snectar behaviour is masked when SSL_OP_ALL is used due to 2223111147Snectar SSL_OP_MICROSOFT_SESS_ID_BUG being set. 2224111147Snectar Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 2225111147Snectar followup to PR #377. 2226111147Snectar [Lutz Jaenicke] 2227111147Snectar 2228111147Snectar *) IA-32 assembler support enhancements: unified ELF targets, support 2229111147Snectar for SCO/Caldera platforms, fix for Cygwin shared build. 2230111147Snectar [Andy Polyakov] 2231111147Snectar 2232111147Snectar *) Add support for FreeBSD on sparc64. As a consequence, support for 2233111147Snectar FreeBSD on non-x86 processors is separate from x86 processors on 2234111147Snectar the config script, much like the NetBSD support. 2235111147Snectar [Richard Levitte & Kris Kennaway <kris@obsecurity.org>] 2236111147Snectar 2237109998Smarkm Changes between 0.9.6h and 0.9.7 [31 Dec 2002] 2238109998Smarkm 2239120631Snectar [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after 2240120631Snectar OpenSSL 0.9.7.] 2241120631Snectar 2242109998Smarkm *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED 2243109998Smarkm code (06) was taken as the first octet of the session ID and the last 2244109998Smarkm octet was ignored consequently. As a result SSLv2 client side session 2245109998Smarkm caching could not have worked due to the session ID mismatch between 2246109998Smarkm client and server. 2247109998Smarkm Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 2248109998Smarkm PR #377. 2249109998Smarkm [Lutz Jaenicke] 2250109998Smarkm 2251109998Smarkm *) Change the declaration of needed Kerberos libraries to use EX_LIBS 2252109998Smarkm instead of the special (and badly supported) LIBKRB5. LIBKRB5 is 2253109998Smarkm removed entirely. 2254109998Smarkm [Richard Levitte] 2255109998Smarkm 2256109998Smarkm *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it 2257109998Smarkm seems that in spite of existing for more than a year, many application 2258109998Smarkm author have done nothing to provide the necessary callbacks, which 2259109998Smarkm means that this particular engine will not work properly anywhere. 2260109998Smarkm This is a very unfortunate situation which forces us, in the name 2261109998Smarkm of usability, to give the hw_ncipher.c a static lock, which is part 2262109998Smarkm of libcrypto. 2263109998Smarkm NOTE: This is for the 0.9.7 series ONLY. This hack will never 2264109998Smarkm appear in 0.9.8 or later. We EXPECT application authors to have 2265109998Smarkm dealt properly with this when 0.9.8 is released (unless we actually 2266109998Smarkm make such changes in the libcrypto locking code that changes will 2267109998Smarkm have to be made anyway). 2268109998Smarkm [Richard Levitte] 2269109998Smarkm 2270109998Smarkm *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content 2271109998Smarkm octets have been read, EOF or an error occurs. Without this change 2272109998Smarkm some truncated ASN1 structures will not produce an error. 2273109998Smarkm [Steve Henson] 2274109998Smarkm 2275109998Smarkm *) Disable Heimdal support, since it hasn't been fully implemented. 2276109998Smarkm Still give the possibility to force the use of Heimdal, but with 2277109998Smarkm warnings and a request that patches get sent to openssl-dev. 2278109998Smarkm [Richard Levitte] 2279109998Smarkm 2280109998Smarkm *) Add the VC-CE target, introduce the WINCE sysname, and add 2281109998Smarkm INSTALL.WCE and appropriate conditionals to make it build. 2282109998Smarkm [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 2283109998Smarkm 2284109998Smarkm *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and 2285109998Smarkm cygssl-x.y.z.dll, where x, y and z are the major, minor and 2286109998Smarkm edit numbers of the version. 2287109998Smarkm [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 2288109998Smarkm 2289109998Smarkm *) Introduce safe string copy and catenation functions 2290109998Smarkm (BUF_strlcpy() and BUF_strlcat()). 2291109998Smarkm [Ben Laurie (CHATS) and Richard Levitte] 2292109998Smarkm 2293109998Smarkm *) Avoid using fixed-size buffers for one-line DNs. 2294109998Smarkm [Ben Laurie (CHATS)] 2295109998Smarkm 2296109998Smarkm *) Add BUF_MEM_grow_clean() to avoid information leakage when 2297109998Smarkm resizing buffers containing secrets, and use where appropriate. 2298109998Smarkm [Ben Laurie (CHATS)] 2299109998Smarkm 2300109998Smarkm *) Avoid using fixed size buffers for configuration file location. 2301109998Smarkm [Ben Laurie (CHATS)] 2302109998Smarkm 2303109998Smarkm *) Avoid filename truncation for various CA files. 2304109998Smarkm [Ben Laurie (CHATS)] 2305109998Smarkm 2306109998Smarkm *) Use sizeof in preference to magic numbers. 2307109998Smarkm [Ben Laurie (CHATS)] 2308109998Smarkm 2309109998Smarkm *) Avoid filename truncation in cert requests. 2310109998Smarkm [Ben Laurie (CHATS)] 2311109998Smarkm 2312109998Smarkm *) Add assertions to check for (supposedly impossible) buffer 2313109998Smarkm overflows. 2314109998Smarkm [Ben Laurie (CHATS)] 2315109998Smarkm 2316109998Smarkm *) Don't cache truncated DNS entries in the local cache (this could 2317109998Smarkm potentially lead to a spoofing attack). 2318109998Smarkm [Ben Laurie (CHATS)] 2319109998Smarkm 2320109998Smarkm *) Fix various buffers to be large enough for hex/decimal 2321109998Smarkm representations in a platform independent manner. 2322109998Smarkm [Ben Laurie (CHATS)] 2323109998Smarkm 2324109998Smarkm *) Add CRYPTO_realloc_clean() to avoid information leakage when 2325109998Smarkm resizing buffers containing secrets, and use where appropriate. 2326109998Smarkm [Ben Laurie (CHATS)] 2327109998Smarkm 2328109998Smarkm *) Add BIO_indent() to avoid much slightly worrying code to do 2329109998Smarkm indents. 2330109998Smarkm [Ben Laurie (CHATS)] 2331109998Smarkm 2332109998Smarkm *) Convert sprintf()/BIO_puts() to BIO_printf(). 2333109998Smarkm [Ben Laurie (CHATS)] 2334109998Smarkm 2335109998Smarkm *) buffer_gets() could terminate with the buffer only half 2336109998Smarkm full. Fixed. 2337109998Smarkm [Ben Laurie (CHATS)] 2338109998Smarkm 2339109998Smarkm *) Add assertions to prevent user-supplied crypto functions from 2340109998Smarkm overflowing internal buffers by having large block sizes, etc. 2341109998Smarkm [Ben Laurie (CHATS)] 2342109998Smarkm 2343109998Smarkm *) New OPENSSL_assert() macro (similar to assert(), but enabled 2344109998Smarkm unconditionally). 2345109998Smarkm [Ben Laurie (CHATS)] 2346109998Smarkm 2347109998Smarkm *) Eliminate unused copy of key in RC4. 2348109998Smarkm [Ben Laurie (CHATS)] 2349109998Smarkm 2350109998Smarkm *) Eliminate unused and incorrectly sized buffers for IV in pem.h. 2351109998Smarkm [Ben Laurie (CHATS)] 2352109998Smarkm 2353109998Smarkm *) Fix off-by-one error in EGD path. 2354109998Smarkm [Ben Laurie (CHATS)] 2355109998Smarkm 2356109998Smarkm *) If RANDFILE path is too long, ignore instead of truncating. 2357109998Smarkm [Ben Laurie (CHATS)] 2358109998Smarkm 2359109998Smarkm *) Eliminate unused and incorrectly sized X.509 structure 2360109998Smarkm CBCParameter. 2361109998Smarkm [Ben Laurie (CHATS)] 2362109998Smarkm 2363109998Smarkm *) Eliminate unused and dangerous function knumber(). 2364109998Smarkm [Ben Laurie (CHATS)] 2365109998Smarkm 2366109998Smarkm *) Eliminate unused and dangerous structure, KSSL_ERR. 2367109998Smarkm [Ben Laurie (CHATS)] 2368109998Smarkm 2369109998Smarkm *) Protect against overlong session ID context length in an encoded 2370109998Smarkm session object. Since these are local, this does not appear to be 2371109998Smarkm exploitable. 2372109998Smarkm [Ben Laurie (CHATS)] 2373109998Smarkm 2374109998Smarkm *) Change from security patch (see 0.9.6e below) that did not affect 2375109998Smarkm the 0.9.6 release series: 2376109998Smarkm 2377109998Smarkm Remote buffer overflow in SSL3 protocol - an attacker could 2378109998Smarkm supply an oversized master key in Kerberos-enabled versions. 2379160814Ssimon (CVE-2002-0657) 2380109998Smarkm [Ben Laurie (CHATS)] 2381109998Smarkm 2382109998Smarkm *) Change the SSL kerb5 codes to match RFC 2712. 2383109998Smarkm [Richard Levitte] 2384109998Smarkm 2385109998Smarkm *) Make -nameopt work fully for req and add -reqopt switch. 2386109998Smarkm [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson] 2387109998Smarkm 2388109998Smarkm *) The "block size" for block ciphers in CFB and OFB mode should be 1. 2389109998Smarkm [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>] 2390109998Smarkm 2391109998Smarkm *) Make sure tests can be performed even if the corresponding algorithms 2392109998Smarkm have been removed entirely. This was also the last step to make 2393109998Smarkm OpenSSL compilable with DJGPP under all reasonable conditions. 2394109998Smarkm [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>] 2395109998Smarkm 2396109998Smarkm *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT 2397109998Smarkm to allow version independent disabling of normally unselected ciphers, 2398109998Smarkm which may be activated as a side-effect of selecting a single cipher. 2399109998Smarkm 2400109998Smarkm (E.g., cipher list string "RSA" enables ciphersuites that are left 2401109998Smarkm out of "ALL" because they do not provide symmetric encryption. 2402109998Smarkm "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.) 2403109998Smarkm [Lutz Jaenicke, Bodo Moeller] 2404109998Smarkm 2405109998Smarkm *) Add appropriate support for separate platform-dependent build 2406109998Smarkm directories. The recommended way to make a platform-dependent 2407109998Smarkm build directory is the following (tested on Linux), maybe with 2408109998Smarkm some local tweaks: 2409109998Smarkm 2410109998Smarkm # Place yourself outside of the OpenSSL source tree. In 2411109998Smarkm # this example, the environment variable OPENSSL_SOURCE 2412109998Smarkm # is assumed to contain the absolute OpenSSL source directory. 2413109998Smarkm mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`" 2414109998Smarkm cd objtree/"`uname -s`-`uname -r`-`uname -m`" 2415111147Snectar (cd $OPENSSL_SOURCE; find . -type f) | while read F; do 2416109998Smarkm mkdir -p `dirname $F` 2417109998Smarkm ln -s $OPENSSL_SOURCE/$F $F 2418109998Smarkm done 2419109998Smarkm 2420109998Smarkm To be absolutely sure not to disturb the source tree, a "make clean" 2421109998Smarkm is a good thing. If it isn't successfull, don't worry about it, 2422109998Smarkm it probably means the source directory is very clean. 2423109998Smarkm [Richard Levitte] 2424109998Smarkm 2425109998Smarkm *) Make sure any ENGINE control commands make local copies of string 2426109998Smarkm pointers passed to them whenever necessary. Otherwise it is possible 2427109998Smarkm the caller may have overwritten (or deallocated) the original string 2428109998Smarkm data when a later ENGINE operation tries to use the stored values. 2429109998Smarkm [G�tz Babin-Ebell <babinebell@trustcenter.de>] 2430109998Smarkm 2431109998Smarkm *) Improve diagnostics in file reading and command-line digests. 2432109998Smarkm [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>] 2433109998Smarkm 2434109998Smarkm *) Add AES modes CFB and OFB to the object database. Correct an 2435109998Smarkm error in AES-CFB decryption. 2436109998Smarkm [Richard Levitte] 2437109998Smarkm 2438109998Smarkm *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this 2439109998Smarkm allows existing EVP_CIPHER_CTX structures to be reused after 2440109998Smarkm calling EVP_*Final(). This behaviour is used by encryption 2441109998Smarkm BIOs and some applications. This has the side effect that 2442109998Smarkm applications must explicitly clean up cipher contexts with 2443109998Smarkm EVP_CIPHER_CTX_cleanup() or they will leak memory. 2444109998Smarkm [Steve Henson] 2445109998Smarkm 2446109998Smarkm *) Check the values of dna and dnb in bn_mul_recursive before calling 2447109998Smarkm bn_mul_comba (a non zero value means the a or b arrays do not contain 2448109998Smarkm n2 elements) and fallback to bn_mul_normal if either is not zero. 2449109998Smarkm [Steve Henson] 2450109998Smarkm 2451109998Smarkm *) Fix escaping of non-ASCII characters when using the -subj option 2452109998Smarkm of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>) 2453109998Smarkm [Lutz Jaenicke] 2454109998Smarkm 2455109998Smarkm *) Make object definitions compliant to LDAP (RFC2256): SN is the short 2456109998Smarkm form for "surname", serialNumber has no short form. 2457109998Smarkm Use "mail" as the short name for "rfc822Mailbox" according to RFC2798; 2458109998Smarkm therefore remove "mail" short name for "internet 7". 2459109998Smarkm The OID for unique identifiers in X509 certificates is 2460109998Smarkm x500UniqueIdentifier, not uniqueIdentifier. 2461109998Smarkm Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>) 2462109998Smarkm [Lutz Jaenicke] 2463109998Smarkm 2464109998Smarkm *) Add an "init" command to the ENGINE config module and auto initialize 2465109998Smarkm ENGINEs. Without any "init" command the ENGINE will be initialized 2466109998Smarkm after all ctrl commands have been executed on it. If init=1 the 2467109998Smarkm ENGINE is initailized at that point (ctrls before that point are run 2468109998Smarkm on the uninitialized ENGINE and after on the initialized one). If 2469109998Smarkm init=0 then the ENGINE will not be iniatialized at all. 2470109998Smarkm [Steve Henson] 2471109998Smarkm 2472109998Smarkm *) Fix the 'app_verify_callback' interface so that the user-defined 2473109998Smarkm argument is actually passed to the callback: In the 2474109998Smarkm SSL_CTX_set_cert_verify_callback() prototype, the callback 2475109998Smarkm declaration has been changed from 2476109998Smarkm int (*cb)() 2477109998Smarkm into 2478109998Smarkm int (*cb)(X509_STORE_CTX *,void *); 2479109998Smarkm in ssl_verify_cert_chain (ssl/ssl_cert.c), the call 2480109998Smarkm i=s->ctx->app_verify_callback(&ctx) 2481109998Smarkm has been changed into 2482109998Smarkm i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg). 2483109998Smarkm 2484109998Smarkm To update applications using SSL_CTX_set_cert_verify_callback(), 2485109998Smarkm a dummy argument can be added to their callback functions. 2486109998Smarkm [D. K. Smetters <smetters@parc.xerox.com>] 2487109998Smarkm 2488109998Smarkm *) Added the '4758cca' ENGINE to support IBM 4758 cards. 2489109998Smarkm [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe] 2490109998Smarkm 2491109998Smarkm *) Add and OPENSSL_LOAD_CONF define which will cause 2492109998Smarkm OpenSSL_add_all_algorithms() to load the openssl.cnf config file. 2493109998Smarkm This allows older applications to transparently support certain 2494109998Smarkm OpenSSL features: such as crypto acceleration and dynamic ENGINE loading. 2495109998Smarkm Two new functions OPENSSL_add_all_algorithms_noconf() which will never 2496109998Smarkm load the config file and OPENSSL_add_all_algorithms_conf() which will 2497109998Smarkm always load it have also been added. 2498109998Smarkm [Steve Henson] 2499109998Smarkm 2500109998Smarkm *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES. 2501109998Smarkm Adjust NIDs and EVP layer. 2502109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 2503109998Smarkm 2504109998Smarkm *) Config modules support in openssl utility. 2505109998Smarkm 2506109998Smarkm Most commands now load modules from the config file, 2507109998Smarkm though in a few (such as version) this isn't done 2508109998Smarkm because it couldn't be used for anything. 2509109998Smarkm 2510109998Smarkm In the case of ca and req the config file used is 2511109998Smarkm the same as the utility itself: that is the -config 2512109998Smarkm command line option can be used to specify an 2513109998Smarkm alternative file. 2514109998Smarkm [Steve Henson] 2515109998Smarkm 2516109998Smarkm *) Move default behaviour from OPENSSL_config(). If appname is NULL 2517109998Smarkm use "openssl_conf" if filename is NULL use default openssl config file. 2518109998Smarkm [Steve Henson] 2519109998Smarkm 2520109998Smarkm *) Add an argument to OPENSSL_config() to allow the use of an alternative 2521109998Smarkm config section name. Add a new flag to tolerate a missing config file 2522109998Smarkm and move code to CONF_modules_load_file(). 2523109998Smarkm [Steve Henson] 2524109998Smarkm 2525109998Smarkm *) Support for crypto accelerator cards from Accelerated Encryption 2526109998Smarkm Processing, www.aep.ie. (Use engine 'aep') 2527109998Smarkm The support was copied from 0.9.6c [engine] and adapted/corrected 2528109998Smarkm to work with the new engine framework. 2529109998Smarkm [AEP Inc. and Richard Levitte] 2530109998Smarkm 2531109998Smarkm *) Support for SureWare crypto accelerator cards from Baltimore 2532109998Smarkm Technologies. (Use engine 'sureware') 2533109998Smarkm The support was copied from 0.9.6c [engine] and adapted 2534109998Smarkm to work with the new engine framework. 2535109998Smarkm [Richard Levitte] 2536109998Smarkm 2537109998Smarkm *) Have the CHIL engine fork-safe (as defined by nCipher) and actually 2538109998Smarkm make the newer ENGINE framework commands for the CHIL engine work. 2539109998Smarkm [Toomas Kiisk <vix@cyber.ee> and Richard Levitte] 2540109998Smarkm 2541109998Smarkm *) Make it possible to produce shared libraries on ReliantUNIX. 2542109998Smarkm [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte] 2543109998Smarkm 2544109998Smarkm *) Add the configuration target debug-linux-ppro. 2545109998Smarkm Make 'openssl rsa' use the general key loading routines 2546109998Smarkm implemented in apps.c, and make those routines able to 2547109998Smarkm handle the key format FORMAT_NETSCAPE and the variant 2548109998Smarkm FORMAT_IISSGC. 2549109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 2550109998Smarkm 2551109998Smarkm *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 2552109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 2553109998Smarkm 2554109998Smarkm *) Add -keyform to rsautl, and document -engine. 2555109998Smarkm [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>] 2556109998Smarkm 2557109998Smarkm *) Change BIO_new_file (crypto/bio/bss_file.c) to use new 2558109998Smarkm BIO_R_NO_SUCH_FILE error code rather than the generic 2559109998Smarkm ERR_R_SYS_LIB error code if fopen() fails with ENOENT. 2560109998Smarkm [Ben Laurie] 2561109998Smarkm 2562109998Smarkm *) Add new functions 2563109998Smarkm ERR_peek_last_error 2564109998Smarkm ERR_peek_last_error_line 2565109998Smarkm ERR_peek_last_error_line_data. 2566109998Smarkm These are similar to 2567109998Smarkm ERR_peek_error 2568109998Smarkm ERR_peek_error_line 2569109998Smarkm ERR_peek_error_line_data, 2570109998Smarkm but report on the latest error recorded rather than the first one 2571109998Smarkm still in the error queue. 2572109998Smarkm [Ben Laurie, Bodo Moeller] 2573109998Smarkm 2574109998Smarkm *) default_algorithms option in ENGINE config module. This allows things 2575109998Smarkm like: 2576109998Smarkm default_algorithms = ALL 2577109998Smarkm default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS 2578109998Smarkm [Steve Henson] 2579109998Smarkm 2580109998Smarkm *) Prelminary ENGINE config module. 2581109998Smarkm [Steve Henson] 2582109998Smarkm 2583109998Smarkm *) New experimental application configuration code. 2584109998Smarkm [Steve Henson] 2585109998Smarkm 2586109998Smarkm *) Change the AES code to follow the same name structure as all other 2587109998Smarkm symmetric ciphers, and behave the same way. Move everything to 2588109998Smarkm the directory crypto/aes, thereby obsoleting crypto/rijndael. 2589109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 2590109998Smarkm 2591109998Smarkm *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c. 2592109998Smarkm [Ben Laurie and Theo de Raadt] 2593109998Smarkm 2594109998Smarkm *) Add option to output public keys in req command. 2595109998Smarkm [Massimiliano Pala madwolf@openca.org] 2596109998Smarkm 2597109998Smarkm *) Use wNAFs in EC_POINTs_mul() for improved efficiency 2598109998Smarkm (up to about 10% better than before for P-192 and P-224). 2599109998Smarkm [Bodo Moeller] 2600109998Smarkm 2601109998Smarkm *) New functions/macros 2602109998Smarkm 2603109998Smarkm SSL_CTX_set_msg_callback(ctx, cb) 2604109998Smarkm SSL_CTX_set_msg_callback_arg(ctx, arg) 2605109998Smarkm SSL_set_msg_callback(ssl, cb) 2606109998Smarkm SSL_set_msg_callback_arg(ssl, arg) 2607109998Smarkm 2608109998Smarkm to request calling a callback function 2609109998Smarkm 2610109998Smarkm void cb(int write_p, int version, int content_type, 2611109998Smarkm const void *buf, size_t len, SSL *ssl, void *arg) 2612109998Smarkm 2613109998Smarkm whenever a protocol message has been completely received 2614109998Smarkm (write_p == 0) or sent (write_p == 1). Here 'version' is the 2615109998Smarkm protocol version according to which the SSL library interprets 2616109998Smarkm the current protocol message (SSL2_VERSION, SSL3_VERSION, or 2617109998Smarkm TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or 2618109998Smarkm the content type as defined in the SSL 3.0/TLS 1.0 protocol 2619109998Smarkm specification (change_cipher_spec(20), alert(21), handshake(22)). 2620109998Smarkm 'buf' and 'len' point to the actual message, 'ssl' to the 2621109998Smarkm SSL object, and 'arg' is the application-defined value set by 2622109998Smarkm SSL[_CTX]_set_msg_callback_arg(). 2623109998Smarkm 2624109998Smarkm 'openssl s_client' and 'openssl s_server' have new '-msg' options 2625109998Smarkm to enable a callback that displays all protocol messages. 2626109998Smarkm [Bodo Moeller] 2627109998Smarkm 2628109998Smarkm *) Change the shared library support so shared libraries are built as 2629109998Smarkm soon as the corresponding static library is finished, and thereby get 2630109998Smarkm openssl and the test programs linked against the shared library. 2631109998Smarkm This still only happens when the keyword "shard" has been given to 2632109998Smarkm the configuration scripts. 2633109998Smarkm 2634109998Smarkm NOTE: shared library support is still an experimental thing, and 2635109998Smarkm backward binary compatibility is still not guaranteed. 2636109998Smarkm ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte] 2637109998Smarkm 2638109998Smarkm *) Add support for Subject Information Access extension. 2639109998Smarkm [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 2640109998Smarkm 2641109998Smarkm *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero 2642109998Smarkm additional bytes when new memory had to be allocated, not just 2643109998Smarkm when reusing an existing buffer. 2644109998Smarkm [Bodo Moeller] 2645109998Smarkm 2646109998Smarkm *) New command line and configuration option 'utf8' for the req command. 2647109998Smarkm This allows field values to be specified as UTF8 strings. 2648109998Smarkm [Steve Henson] 2649109998Smarkm 2650109998Smarkm *) Add -multi and -mr options to "openssl speed" - giving multiple parallel 2651109998Smarkm runs for the former and machine-readable output for the latter. 2652109998Smarkm [Ben Laurie] 2653109998Smarkm 2654109998Smarkm *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion 2655109998Smarkm of the e-mail address in the DN (i.e., it will go into a certificate 2656109998Smarkm extension only). The new configuration file option 'email_in_dn = no' 2657109998Smarkm has the same effect. 2658109998Smarkm [Massimiliano Pala madwolf@openca.org] 2659109998Smarkm 2660109998Smarkm *) Change all functions with names starting with des_ to be starting 2661109998Smarkm with DES_ instead. Add wrappers that are compatible with libdes, 2662109998Smarkm but are named _ossl_old_des_*. Finally, add macros that map the 2663109998Smarkm des_* symbols to the corresponding _ossl_old_des_* if libdes 2664109998Smarkm compatibility is desired. If OpenSSL 0.9.6c compatibility is 2665109998Smarkm desired, the des_* symbols will be mapped to DES_*, with one 2666109998Smarkm exception. 2667109998Smarkm 2668109998Smarkm Since we provide two compatibility mappings, the user needs to 2669109998Smarkm define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes 2670109998Smarkm compatibility is desired. The default (i.e., when that macro 2671109998Smarkm isn't defined) is OpenSSL 0.9.6c compatibility. 2672109998Smarkm 2673109998Smarkm There are also macros that enable and disable the support of old 2674109998Smarkm des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT 2675109998Smarkm and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those 2676109998Smarkm are defined, the default will apply: to support the old des routines. 2677109998Smarkm 2678109998Smarkm In either case, one must include openssl/des.h to get the correct 2679109998Smarkm definitions. Do not try to just include openssl/des_old.h, that 2680109998Smarkm won't work. 2681109998Smarkm 2682109998Smarkm NOTE: This is a major break of an old API into a new one. Software 2683109998Smarkm authors are encouraged to switch to the DES_ style functions. Some 2684109998Smarkm time in the future, des_old.h and the libdes compatibility functions 2685109998Smarkm will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the 2686109998Smarkm default), and then completely removed. 2687109998Smarkm [Richard Levitte] 2688109998Smarkm 2689109998Smarkm *) Test for certificates which contain unsupported critical extensions. 2690109998Smarkm If such a certificate is found during a verify operation it is 2691109998Smarkm rejected by default: this behaviour can be overridden by either 2692109998Smarkm handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or 2693109998Smarkm by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function 2694109998Smarkm X509_supported_extension() has also been added which returns 1 if a 2695109998Smarkm particular extension is supported. 2696109998Smarkm [Steve Henson] 2697109998Smarkm 2698109998Smarkm *) Modify the behaviour of EVP cipher functions in similar way to digests 2699109998Smarkm to retain compatibility with existing code. 2700109998Smarkm [Steve Henson] 2701109998Smarkm 2702109998Smarkm *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain 2703109998Smarkm compatibility with existing code. In particular the 'ctx' parameter does 2704109998Smarkm not have to be to be initialized before the call to EVP_DigestInit() and 2705109998Smarkm it is tidied up after a call to EVP_DigestFinal(). New function 2706109998Smarkm EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function 2707109998Smarkm EVP_MD_CTX_copy() changed to not require the destination to be 2708109998Smarkm initialized valid and new function EVP_MD_CTX_copy_ex() added which 2709109998Smarkm requires the destination to be valid. 2710109998Smarkm 2711109998Smarkm Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(), 2712109998Smarkm EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex(). 2713109998Smarkm [Steve Henson] 2714109998Smarkm 2715109998Smarkm *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it 2716109998Smarkm so that complete 'Handshake' protocol structures are kept in memory 2717109998Smarkm instead of overwriting 'msg_type' and 'length' with 'body' data. 2718109998Smarkm [Bodo Moeller] 2719109998Smarkm 2720109998Smarkm *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32. 2721109998Smarkm [Massimo Santin via Richard Levitte] 2722109998Smarkm 2723109998Smarkm *) Major restructuring to the underlying ENGINE code. This includes 2724109998Smarkm reduction of linker bloat, separation of pure "ENGINE" manipulation 2725109998Smarkm (initialisation, etc) from functionality dealing with implementations 2726109998Smarkm of specific crypto iterfaces. This change also introduces integrated 2727109998Smarkm support for symmetric ciphers and digest implementations - so ENGINEs 2728109998Smarkm can now accelerate these by providing EVP_CIPHER and EVP_MD 2729109998Smarkm implementations of their own. This is detailed in crypto/engine/README 2730109998Smarkm as it couldn't be adequately described here. However, there are a few 2731109998Smarkm API changes worth noting - some RSA, DSA, DH, and RAND functions that 2732109998Smarkm were changed in the original introduction of ENGINE code have now 2733109998Smarkm reverted back - the hooking from this code to ENGINE is now a good 2734109998Smarkm deal more passive and at run-time, operations deal directly with 2735109998Smarkm RSA_METHODs, DSA_METHODs (etc) as they did before, rather than 2736109998Smarkm dereferencing through an ENGINE pointer any more. Also, the ENGINE 2737109998Smarkm functions dealing with BN_MOD_EXP[_CRT] handlers have been removed - 2738109998Smarkm they were not being used by the framework as there is no concept of a 2739109998Smarkm BIGNUM_METHOD and they could not be generalised to the new 2740109998Smarkm 'ENGINE_TABLE' mechanism that underlies the new code. Similarly, 2741109998Smarkm ENGINE_cpy() has been removed as it cannot be consistently defined in 2742109998Smarkm the new code. 2743109998Smarkm [Geoff Thorpe] 2744109998Smarkm 2745109998Smarkm *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds. 2746109998Smarkm [Steve Henson] 2747109998Smarkm 2748109998Smarkm *) Change mkdef.pl to sort symbols that get the same entry number, 2749109998Smarkm and make sure the automatically generated functions ERR_load_* 2750109998Smarkm become part of libeay.num as well. 2751109998Smarkm [Richard Levitte] 2752109998Smarkm 2753109998Smarkm *) New function SSL_renegotiate_pending(). This returns true once 2754109998Smarkm renegotiation has been requested (either SSL_renegotiate() call 2755109998Smarkm or HelloRequest/ClientHello receveived from the peer) and becomes 2756109998Smarkm false once a handshake has been completed. 2757109998Smarkm (For servers, SSL_renegotiate() followed by SSL_do_handshake() 2758109998Smarkm sends a HelloRequest, but does not ensure that a handshake takes 2759109998Smarkm place. SSL_renegotiate_pending() is useful for checking if the 2760109998Smarkm client has followed the request.) 2761109998Smarkm [Bodo Moeller] 2762109998Smarkm 2763109998Smarkm *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION. 2764109998Smarkm By default, clients may request session resumption even during 2765109998Smarkm renegotiation (if session ID contexts permit); with this option, 2766109998Smarkm session resumption is possible only in the first handshake. 2767109998Smarkm 2768109998Smarkm SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes 2769109998Smarkm more bits available for options that should not be part of 2770109998Smarkm SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION). 2771109998Smarkm [Bodo Moeller] 2772109998Smarkm 2773109998Smarkm *) Add some demos for certificate and certificate request creation. 2774109998Smarkm [Steve Henson] 2775109998Smarkm 2776109998Smarkm *) Make maximum certificate chain size accepted from the peer application 2777109998Smarkm settable (SSL*_get/set_max_cert_list()), as proposed by 2778109998Smarkm "Douglas E. Engert" <deengert@anl.gov>. 2779109998Smarkm [Lutz Jaenicke] 2780109998Smarkm 2781109998Smarkm *) Add support for shared libraries for Unixware-7 2782109998Smarkm (Boyd Lynn Gerber <gerberb@zenez.com>). 2783109998Smarkm [Lutz Jaenicke] 2784109998Smarkm 2785109998Smarkm *) Add a "destroy" handler to ENGINEs that allows structural cleanup to 2786109998Smarkm be done prior to destruction. Use this to unload error strings from 2787109998Smarkm ENGINEs that load their own error strings. NB: This adds two new API 2788109998Smarkm functions to "get" and "set" this destroy handler in an ENGINE. 2789109998Smarkm [Geoff Thorpe] 2790109998Smarkm 2791109998Smarkm *) Alter all existing ENGINE implementations (except "openssl" and 2792109998Smarkm "openbsd") to dynamically instantiate their own error strings. This 2793109998Smarkm makes them more flexible to be built both as statically-linked ENGINEs 2794109998Smarkm and self-contained shared-libraries loadable via the "dynamic" ENGINE. 2795109998Smarkm Also, add stub code to each that makes building them as self-contained 2796109998Smarkm shared-libraries easier (see README.ENGINE). 2797109998Smarkm [Geoff Thorpe] 2798109998Smarkm 2799109998Smarkm *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE 2800109998Smarkm implementations into applications that are completely implemented in 2801109998Smarkm self-contained shared-libraries. The "dynamic" ENGINE exposes control 2802109998Smarkm commands that can be used to configure what shared-library to load and 2803109998Smarkm to control aspects of the way it is handled. Also, made an update to 2804109998Smarkm the README.ENGINE file that brings its information up-to-date and 2805109998Smarkm provides some information and instructions on the "dynamic" ENGINE 2806109998Smarkm (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc). 2807109998Smarkm [Geoff Thorpe] 2808109998Smarkm 2809109998Smarkm *) Make it possible to unload ranges of ERR strings with a new 2810109998Smarkm "ERR_unload_strings" function. 2811109998Smarkm [Geoff Thorpe] 2812109998Smarkm 2813109998Smarkm *) Add a copy() function to EVP_MD. 2814109998Smarkm [Ben Laurie] 2815109998Smarkm 2816109998Smarkm *) Make EVP_MD routines take a context pointer instead of just the 2817109998Smarkm md_data void pointer. 2818109998Smarkm [Ben Laurie] 2819109998Smarkm 2820109998Smarkm *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates 2821109998Smarkm that the digest can only process a single chunk of data 2822109998Smarkm (typically because it is provided by a piece of 2823109998Smarkm hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application 2824109998Smarkm is only going to provide a single chunk of data, and hence the 2825109998Smarkm framework needn't accumulate the data for oneshot drivers. 2826109998Smarkm [Ben Laurie] 2827109998Smarkm 2828109998Smarkm *) As with "ERR", make it possible to replace the underlying "ex_data" 2829109998Smarkm functions. This change also alters the storage and management of global 2830109998Smarkm ex_data state - it's now all inside ex_data.c and all "class" code (eg. 2831109998Smarkm RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class 2832109998Smarkm index counters. The API functions that use this state have been changed 2833109998Smarkm to take a "class_index" rather than pointers to the class's local STACK 2834109998Smarkm and counter, and there is now an API function to dynamically create new 2835109998Smarkm classes. This centralisation allows us to (a) plug a lot of the 2836109998Smarkm thread-safety problems that existed, and (b) makes it possible to clean 2837109998Smarkm up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b) 2838109998Smarkm such data would previously have always leaked in application code and 2839109998Smarkm workarounds were in place to make the memory debugging turn a blind eye 2840109998Smarkm to it. Application code that doesn't use this new function will still 2841109998Smarkm leak as before, but their memory debugging output will announce it now 2842109998Smarkm rather than letting it slide. 2843109998Smarkm 2844109998Smarkm Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change 2845109998Smarkm induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now 2846109998Smarkm has a return value to indicate success or failure. 2847109998Smarkm [Geoff Thorpe] 2848109998Smarkm 2849109998Smarkm *) Make it possible to replace the underlying "ERR" functions such that the 2850109998Smarkm global state (2 LHASH tables and 2 locks) is only used by the "default" 2851109998Smarkm implementation. This change also adds two functions to "get" and "set" 2852109998Smarkm the implementation prior to it being automatically set the first time 2853109998Smarkm any other ERR function takes place. Ie. an application can call "get", 2854109998Smarkm pass the return value to a module it has just loaded, and that module 2855109998Smarkm can call its own "set" function using that value. This means the 2856109998Smarkm module's "ERR" operations will use (and modify) the error state in the 2857109998Smarkm application and not in its own statically linked copy of OpenSSL code. 2858109998Smarkm [Geoff Thorpe] 2859109998Smarkm 2860109998Smarkm *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment 2861109998Smarkm reference counts. This performs normal REF_PRINT/REF_CHECK macros on 2862109998Smarkm the operation, and provides a more encapsulated way for external code 2863109998Smarkm (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code 2864109998Smarkm to use these functions rather than manually incrementing the counts. 2865109998Smarkm 2866109998Smarkm Also rename "DSO_up()" function to more descriptive "DSO_up_ref()". 2867109998Smarkm [Geoff Thorpe] 2868109998Smarkm 2869109998Smarkm *) Add EVP test program. 2870109998Smarkm [Ben Laurie] 2871109998Smarkm 2872109998Smarkm *) Add symmetric cipher support to ENGINE. Expect the API to change! 2873109998Smarkm [Ben Laurie] 2874109998Smarkm 2875109998Smarkm *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name() 2876109998Smarkm X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(), 2877109998Smarkm X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate(). 2878109998Smarkm These allow a CRL to be built without having to access X509_CRL fields 2879109998Smarkm directly. Modify 'ca' application to use new functions. 2880109998Smarkm [Steve Henson] 2881109998Smarkm 2882109998Smarkm *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended 2883109998Smarkm bug workarounds. Rollback attack detection is a security feature. 2884109998Smarkm The problem will only arise on OpenSSL servers when TLSv1 is not 2885109998Smarkm available (sslv3_server_method() or SSL_OP_NO_TLSv1). 2886109998Smarkm Software authors not wanting to support TLSv1 will have special reasons 2887109998Smarkm for their choice and can explicitly enable this option. 2888109998Smarkm [Bodo Moeller, Lutz Jaenicke] 2889109998Smarkm 2890109998Smarkm *) Rationalise EVP so it can be extended: don't include a union of 2891109998Smarkm cipher/digest structures, add init/cleanup functions for EVP_MD_CTX 2892109998Smarkm (similar to those existing for EVP_CIPHER_CTX). 2893109998Smarkm Usage example: 2894109998Smarkm 2895109998Smarkm EVP_MD_CTX md; 2896109998Smarkm 2897109998Smarkm EVP_MD_CTX_init(&md); /* new function call */ 2898109998Smarkm EVP_DigestInit(&md, EVP_sha1()); 2899109998Smarkm EVP_DigestUpdate(&md, in, len); 2900109998Smarkm EVP_DigestFinal(&md, out, NULL); 2901109998Smarkm EVP_MD_CTX_cleanup(&md); /* new function call */ 2902109998Smarkm 2903109998Smarkm [Ben Laurie] 2904109998Smarkm 2905109998Smarkm *) Make DES key schedule conform to the usual scheme, as well as 2906109998Smarkm correcting its structure. This means that calls to DES functions 2907109998Smarkm now have to pass a pointer to a des_key_schedule instead of a 2908109998Smarkm plain des_key_schedule (which was actually always a pointer 2909109998Smarkm anyway): E.g., 2910109998Smarkm 2911109998Smarkm des_key_schedule ks; 2912109998Smarkm 2913109998Smarkm des_set_key_checked(..., &ks); 2914109998Smarkm des_ncbc_encrypt(..., &ks, ...); 2915109998Smarkm 2916109998Smarkm (Note that a later change renames 'des_...' into 'DES_...'.) 2917109998Smarkm [Ben Laurie] 2918109998Smarkm 2919109998Smarkm *) Initial reduction of linker bloat: the use of some functions, such as 2920109998Smarkm PEM causes large amounts of unused functions to be linked in due to 2921109998Smarkm poor organisation. For example pem_all.c contains every PEM function 2922109998Smarkm which has a knock on effect of linking in large amounts of (unused) 2923109998Smarkm ASN1 code. Grouping together similar functions and splitting unrelated 2924109998Smarkm functions prevents this. 2925109998Smarkm [Steve Henson] 2926109998Smarkm 2927109998Smarkm *) Cleanup of EVP macros. 2928109998Smarkm [Ben Laurie] 2929109998Smarkm 2930109998Smarkm *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the 2931109998Smarkm correct _ecb suffix. 2932109998Smarkm [Ben Laurie] 2933109998Smarkm 2934109998Smarkm *) Add initial OCSP responder support to ocsp application. The 2935109998Smarkm revocation information is handled using the text based index 2936109998Smarkm use by the ca application. The responder can either handle 2937109998Smarkm requests generated internally, supplied in files (for example 2938109998Smarkm via a CGI script) or using an internal minimal server. 2939109998Smarkm [Steve Henson] 2940109998Smarkm 2941109998Smarkm *) Add configuration choices to get zlib compression for TLS. 2942109998Smarkm [Richard Levitte] 2943109998Smarkm 2944109998Smarkm *) Changes to Kerberos SSL for RFC 2712 compliance: 2945109998Smarkm 1. Implemented real KerberosWrapper, instead of just using 2946109998Smarkm KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>] 2947109998Smarkm 2. Implemented optional authenticator field of KerberosWrapper. 2948109998Smarkm 2949109998Smarkm Added openssl-style ASN.1 macros for Kerberos ticket, ap_req, 2950109998Smarkm and authenticator structs; see crypto/krb5/. 2951109998Smarkm 2952109998Smarkm Generalized Kerberos calls to support multiple Kerberos libraries. 2953109998Smarkm [Vern Staats <staatsvr@asc.hpc.mil>, 2954109998Smarkm Jeffrey Altman <jaltman@columbia.edu> 2955109998Smarkm via Richard Levitte] 2956109998Smarkm 2957109998Smarkm *) Cause 'openssl speed' to use fully hard-coded DSA keys as it 2958109998Smarkm already does with RSA. testdsa.h now has 'priv_key/pub_key' 2959109998Smarkm values for each of the key sizes rather than having just 2960109998Smarkm parameters (and 'speed' generating keys each time). 2961109998Smarkm [Geoff Thorpe] 2962109998Smarkm 2963109998Smarkm *) Speed up EVP routines. 2964109998Smarkm Before: 2965109998Smarkmencrypt 2966109998Smarkmtype 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes 2967109998Smarkmdes-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k 2968109998Smarkmdes-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k 2969109998Smarkmdes-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k 2970109998Smarkmdecrypt 2971109998Smarkmdes-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k 2972109998Smarkmdes-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k 2973109998Smarkmdes-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k 2974109998Smarkm After: 2975109998Smarkmencrypt 2976109998Smarkmdes-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k 2977109998Smarkmdecrypt 2978109998Smarkmdes-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k 2979109998Smarkm [Ben Laurie] 2980109998Smarkm 2981109998Smarkm *) Added the OS2-EMX target. 2982109998Smarkm ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte] 2983109998Smarkm 2984109998Smarkm *) Rewrite apps to use NCONF routines instead of the old CONF. New functions 2985109998Smarkm to support NCONF routines in extension code. New function CONF_set_nconf() 2986109998Smarkm to allow functions which take an NCONF to also handle the old LHASH 2987109998Smarkm structure: this means that the old CONF compatible routines can be 2988109998Smarkm retained (in particular wrt extensions) without having to duplicate the 2989109998Smarkm code. New function X509V3_add_ext_nconf_sk to add extensions to a stack. 2990109998Smarkm [Steve Henson] 2991109998Smarkm 2992109998Smarkm *) Enhance the general user interface with mechanisms for inner control 2993109998Smarkm and with possibilities to have yes/no kind of prompts. 2994109998Smarkm [Richard Levitte] 2995109998Smarkm 2996109998Smarkm *) Change all calls to low level digest routines in the library and 2997109998Smarkm applications to use EVP. Add missing calls to HMAC_cleanup() and 2998109998Smarkm don't assume HMAC_CTX can be copied using memcpy(). 2999109998Smarkm [Verdon Walker <VWalker@novell.com>, Steve Henson] 3000109998Smarkm 3001109998Smarkm *) Add the possibility to control engines through control names but with 3002109998Smarkm arbitrary arguments instead of just a string. 3003109998Smarkm Change the key loaders to take a UI_METHOD instead of a callback 3004109998Smarkm function pointer. NOTE: this breaks binary compatibility with earlier 3005109998Smarkm versions of OpenSSL [engine]. 3006109998Smarkm Adapt the nCipher code for these new conditions and add a card insertion 3007109998Smarkm callback. 3008109998Smarkm [Richard Levitte] 3009109998Smarkm 3010109998Smarkm *) Enhance the general user interface with mechanisms to better support 3011109998Smarkm dialog box interfaces, application-defined prompts, the possibility 3012109998Smarkm to use defaults (for example default passwords from somewhere else) 3013109998Smarkm and interrupts/cancellations. 3014109998Smarkm [Richard Levitte] 3015109998Smarkm 3016109998Smarkm *) Tidy up PKCS#12 attribute handling. Add support for the CSP name 3017109998Smarkm attribute in PKCS#12 files, add new -CSP option to pkcs12 utility. 3018109998Smarkm [Steve Henson] 3019109998Smarkm 3020109998Smarkm *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also 3021109998Smarkm tidy up some unnecessarily weird code in 'sk_new()'). 3022109998Smarkm [Geoff, reported by Diego Tartara <dtartara@novamens.com>] 3023109998Smarkm 3024109998Smarkm *) Change the key loading routines for ENGINEs to use the same kind 3025109998Smarkm callback (pem_password_cb) as all other routines that need this 3026109998Smarkm kind of callback. 3027109998Smarkm [Richard Levitte] 3028109998Smarkm 3029109998Smarkm *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with 3030109998Smarkm 256 bit (=32 byte) keys. Of course seeding with more entropy bytes 3031109998Smarkm than this minimum value is recommended. 3032109998Smarkm [Lutz Jaenicke] 3033109998Smarkm 3034109998Smarkm *) New random seeder for OpenVMS, using the system process statistics 3035109998Smarkm that are easily reachable. 3036109998Smarkm [Richard Levitte] 3037109998Smarkm 3038109998Smarkm *) Windows apparently can't transparently handle global 3039109998Smarkm variables defined in DLLs. Initialisations such as: 3040109998Smarkm 3041109998Smarkm const ASN1_ITEM *it = &ASN1_INTEGER_it; 3042109998Smarkm 3043109998Smarkm wont compile. This is used by the any applications that need to 3044109998Smarkm declare their own ASN1 modules. This was fixed by adding the option 3045109998Smarkm EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly 3046109998Smarkm needed for static libraries under Win32. 3047109998Smarkm [Steve Henson] 3048109998Smarkm 3049109998Smarkm *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle 3050109998Smarkm setting of purpose and trust fields. New X509_STORE trust and 3051109998Smarkm purpose functions and tidy up setting in other SSL functions. 3052109998Smarkm [Steve Henson] 3053109998Smarkm 3054109998Smarkm *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE 3055109998Smarkm structure. These are inherited by X509_STORE_CTX when it is 3056109998Smarkm initialised. This allows various defaults to be set in the 3057109998Smarkm X509_STORE structure (such as flags for CRL checking and custom 3058109998Smarkm purpose or trust settings) for functions which only use X509_STORE_CTX 3059109998Smarkm internally such as S/MIME. 3060109998Smarkm 3061109998Smarkm Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and 3062109998Smarkm trust settings if they are not set in X509_STORE. This allows X509_STORE 3063109998Smarkm purposes and trust (in S/MIME for example) to override any set by default. 3064109998Smarkm 3065109998Smarkm Add command line options for CRL checking to smime, s_client and s_server 3066109998Smarkm applications. 3067109998Smarkm [Steve Henson] 3068109998Smarkm 3069109998Smarkm *) Initial CRL based revocation checking. If the CRL checking flag(s) 3070109998Smarkm are set then the CRL is looked up in the X509_STORE structure and 3071109998Smarkm its validity and signature checked, then if the certificate is found 3072109998Smarkm in the CRL the verify fails with a revoked error. 3073109998Smarkm 3074109998Smarkm Various new CRL related callbacks added to X509_STORE_CTX structure. 3075109998Smarkm 3076109998Smarkm Command line options added to 'verify' application to support this. 3077109998Smarkm 3078109998Smarkm This needs some additional work, such as being able to handle multiple 3079109998Smarkm CRLs with different times, extension based lookup (rather than just 3080109998Smarkm by subject name) and ultimately more complete V2 CRL extension 3081109998Smarkm handling. 3082109998Smarkm [Steve Henson] 3083109998Smarkm 3084109998Smarkm *) Add a general user interface API (crypto/ui/). This is designed 3085109998Smarkm to replace things like des_read_password and friends (backward 3086109998Smarkm compatibility functions using this new API are provided). 3087109998Smarkm The purpose is to remove prompting functions from the DES code 3088109998Smarkm section as well as provide for prompting through dialog boxes in 3089109998Smarkm a window system and the like. 3090109998Smarkm [Richard Levitte] 3091109998Smarkm 3092109998Smarkm *) Add "ex_data" support to ENGINE so implementations can add state at a 3093109998Smarkm per-structure level rather than having to store it globally. 3094109998Smarkm [Geoff] 3095109998Smarkm 3096109998Smarkm *) Make it possible for ENGINE structures to be copied when retrieved by 3097109998Smarkm ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY. 3098109998Smarkm This causes the "original" ENGINE structure to act like a template, 3099109998Smarkm analogous to the RSA vs. RSA_METHOD type of separation. Because of this 3100109998Smarkm operational state can be localised to each ENGINE structure, despite the 3101109998Smarkm fact they all share the same "methods". New ENGINE structures returned in 3102109998Smarkm this case have no functional references and the return value is the single 3103109998Smarkm structural reference. This matches the single structural reference returned 3104109998Smarkm by ENGINE_by_id() normally, when it is incremented on the pre-existing 3105109998Smarkm ENGINE structure. 3106109998Smarkm [Geoff] 3107109998Smarkm 3108109998Smarkm *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this 3109109998Smarkm needs to match any other type at all we need to manually clear the 3110109998Smarkm tag cache. 3111109998Smarkm [Steve Henson] 3112109998Smarkm 3113109998Smarkm *) Changes to the "openssl engine" utility to include; 3114109998Smarkm - verbosity levels ('-v', '-vv', and '-vvv') that provide information 3115109998Smarkm about an ENGINE's available control commands. 3116109998Smarkm - executing control commands from command line arguments using the 3117109998Smarkm '-pre' and '-post' switches. '-post' is only used if '-t' is 3118109998Smarkm specified and the ENGINE is successfully initialised. The syntax for 3119109998Smarkm the individual commands are colon-separated, for example; 3120109998Smarkm openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so 3121109998Smarkm [Geoff] 3122109998Smarkm 3123109998Smarkm *) New dynamic control command support for ENGINEs. ENGINEs can now 3124109998Smarkm declare their own commands (numbers), names (strings), descriptions, 3125109998Smarkm and input types for run-time discovery by calling applications. A 3126109998Smarkm subset of these commands are implicitly classed as "executable" 3127109998Smarkm depending on their input type, and only these can be invoked through 3128109998Smarkm the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this 3129109998Smarkm can be based on user input, config files, etc). The distinction is 3130109998Smarkm that "executable" commands cannot return anything other than a boolean 3131109998Smarkm result and can only support numeric or string input, whereas some 3132109998Smarkm discoverable commands may only be for direct use through 3133109998Smarkm ENGINE_ctrl(), eg. supporting the exchange of binary data, function 3134109998Smarkm pointers, or other custom uses. The "executable" commands are to 3135109998Smarkm support parameterisations of ENGINE behaviour that can be 3136109998Smarkm unambiguously defined by ENGINEs and used consistently across any 3137109998Smarkm OpenSSL-based application. Commands have been added to all the 3138109998Smarkm existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow 3139109998Smarkm control over shared-library paths without source code alterations. 3140109998Smarkm [Geoff] 3141109998Smarkm 3142109998Smarkm *) Changed all ENGINE implementations to dynamically allocate their 3143109998Smarkm ENGINEs rather than declaring them statically. Apart from this being 3144109998Smarkm necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction, 3145109998Smarkm this also allows the implementations to compile without using the 3146109998Smarkm internal engine_int.h header. 3147109998Smarkm [Geoff] 3148109998Smarkm 3149109998Smarkm *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a 3150109998Smarkm 'const' value. Any code that should be able to modify a RAND_METHOD 3151109998Smarkm should already have non-const pointers to it (ie. they should only 3152109998Smarkm modify their own ones). 3153109998Smarkm [Geoff] 3154109998Smarkm 3155109998Smarkm *) Made a variety of little tweaks to the ENGINE code. 3156109998Smarkm - "atalla" and "ubsec" string definitions were moved from header files 3157109998Smarkm to C code. "nuron" string definitions were placed in variables 3158109998Smarkm rather than hard-coded - allowing parameterisation of these values 3159109998Smarkm later on via ctrl() commands. 3160109998Smarkm - Removed unused "#if 0"'d code. 3161109998Smarkm - Fixed engine list iteration code so it uses ENGINE_free() to release 3162109998Smarkm structural references. 3163109998Smarkm - Constified the RAND_METHOD element of ENGINE structures. 3164109998Smarkm - Constified various get/set functions as appropriate and added 3165109998Smarkm missing functions (including a catch-all ENGINE_cpy that duplicates 3166109998Smarkm all ENGINE values onto a new ENGINE except reference counts/state). 3167109998Smarkm - Removed NULL parameter checks in get/set functions. Setting a method 3168109998Smarkm or function to NULL is a way of cancelling out a previously set 3169109998Smarkm value. Passing a NULL ENGINE parameter is just plain stupid anyway 3170109998Smarkm and doesn't justify the extra error symbols and code. 3171109998Smarkm - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for 3172109998Smarkm flags from engine_int.h to engine.h. 3173109998Smarkm - Changed prototypes for ENGINE handler functions (init(), finish(), 3174109998Smarkm ctrl(), key-load functions, etc) to take an (ENGINE*) parameter. 3175109998Smarkm [Geoff] 3176109998Smarkm 3177109998Smarkm *) Implement binary inversion algorithm for BN_mod_inverse in addition 3178109998Smarkm to the algorithm using long division. The binary algorithm can be 3179109998Smarkm used only if the modulus is odd. On 32-bit systems, it is faster 3180109998Smarkm only for relatively small moduli (roughly 20-30% for 128-bit moduli, 3181109998Smarkm roughly 5-15% for 256-bit moduli), so we use it only for moduli 3182109998Smarkm up to 450 bits. In 64-bit environments, the binary algorithm 3183109998Smarkm appears to be advantageous for much longer moduli; here we use it 3184109998Smarkm for moduli up to 2048 bits. 3185109998Smarkm [Bodo Moeller] 3186109998Smarkm 3187109998Smarkm *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code 3188109998Smarkm could not support the combine flag in choice fields. 3189109998Smarkm [Steve Henson] 3190109998Smarkm 3191109998Smarkm *) Add a 'copy_extensions' option to the 'ca' utility. This copies 3192109998Smarkm extensions from a certificate request to the certificate. 3193109998Smarkm [Steve Henson] 3194109998Smarkm 3195109998Smarkm *) Allow multiple 'certopt' and 'nameopt' options to be separated 3196109998Smarkm by commas. Add 'namopt' and 'certopt' options to the 'ca' config 3197109998Smarkm file: this allows the display of the certificate about to be 3198109998Smarkm signed to be customised, to allow certain fields to be included 3199109998Smarkm or excluded and extension details. The old system didn't display 3200109998Smarkm multicharacter strings properly, omitted fields not in the policy 3201109998Smarkm and couldn't display additional details such as extensions. 3202109998Smarkm [Steve Henson] 3203109998Smarkm 3204109998Smarkm *) Function EC_POINTs_mul for multiple scalar multiplication 3205109998Smarkm of an arbitrary number of elliptic curve points 3206109998Smarkm \sum scalars[i]*points[i], 3207109998Smarkm optionally including the generator defined for the EC_GROUP: 3208109998Smarkm scalar*generator + \sum scalars[i]*points[i]. 3209109998Smarkm 3210109998Smarkm EC_POINT_mul is a simple wrapper function for the typical case 3211109998Smarkm that the point list has just one item (besides the optional 3212109998Smarkm generator). 3213109998Smarkm [Bodo Moeller] 3214109998Smarkm 3215109998Smarkm *) First EC_METHODs for curves over GF(p): 3216109998Smarkm 3217109998Smarkm EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr 3218109998Smarkm operations and provides various method functions that can also 3219109998Smarkm operate with faster implementations of modular arithmetic. 3220109998Smarkm 3221109998Smarkm EC_GFp_mont_method() reuses most functions that are part of 3222109998Smarkm EC_GFp_simple_method, but uses Montgomery arithmetic. 3223109998Smarkm 3224109998Smarkm [Bodo Moeller; point addition and point doubling 3225109998Smarkm implementation directly derived from source code provided by 3226109998Smarkm Lenka Fibikova <fibikova@exp-math.uni-essen.de>] 3227109998Smarkm 3228109998Smarkm *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h, 3229109998Smarkm crypto/ec/ec_lib.c): 3230109998Smarkm 3231109998Smarkm Curves are EC_GROUP objects (with an optional group generator) 3232109998Smarkm based on EC_METHODs that are built into the library. 3233109998Smarkm 3234109998Smarkm Points are EC_POINT objects based on EC_GROUP objects. 3235109998Smarkm 3236109998Smarkm Most of the framework would be able to handle curves over arbitrary 3237109998Smarkm finite fields, but as there are no obvious types for fields other 3238109998Smarkm than GF(p), some functions are limited to that for now. 3239109998Smarkm [Bodo Moeller] 3240109998Smarkm 3241109998Smarkm *) Add the -HTTP option to s_server. It is similar to -WWW, but requires 3242109998Smarkm that the file contains a complete HTTP response. 3243109998Smarkm [Richard Levitte] 3244109998Smarkm 3245109998Smarkm *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl 3246109998Smarkm change the def and num file printf format specifier from "%-40sXXX" 3247109998Smarkm to "%-39s XXX". The latter will always guarantee a space after the 3248109998Smarkm field while the former will cause them to run together if the field 3249109998Smarkm is 40 of more characters long. 3250109998Smarkm [Steve Henson] 3251109998Smarkm 3252109998Smarkm *) Constify the cipher and digest 'method' functions and structures 3253109998Smarkm and modify related functions to take constant EVP_MD and EVP_CIPHER 3254109998Smarkm pointers. 3255109998Smarkm [Steve Henson] 3256109998Smarkm 3257109998Smarkm *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them 3258109998Smarkm in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32. 3259109998Smarkm [Bodo Moeller] 3260109998Smarkm 3261109998Smarkm *) Modify EVP_Digest*() routines so they now return values. Although the 3262109998Smarkm internal software routines can never fail additional hardware versions 3263109998Smarkm might. 3264109998Smarkm [Steve Henson] 3265109998Smarkm 3266109998Smarkm *) Clean up crypto/err/err.h and change some error codes to avoid conflicts: 3267109998Smarkm 3268109998Smarkm Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7 3269109998Smarkm (= ERR_R_PKCS7_LIB); it is now 64 instead of 32. 3270109998Smarkm 3271109998Smarkm ASN1 error codes 3272109998Smarkm ERR_R_NESTED_ASN1_ERROR 3273109998Smarkm ... 3274109998Smarkm ERR_R_MISSING_ASN1_EOS 3275109998Smarkm were 4 .. 9, conflicting with 3276109998Smarkm ERR_LIB_RSA (= ERR_R_RSA_LIB) 3277109998Smarkm ... 3278109998Smarkm ERR_LIB_PEM (= ERR_R_PEM_LIB). 3279109998Smarkm They are now 58 .. 63 (i.e., just below ERR_R_FATAL). 3280109998Smarkm 3281109998Smarkm Add new error code 'ERR_R_INTERNAL_ERROR'. 3282109998Smarkm [Bodo Moeller] 3283109998Smarkm 3284109998Smarkm *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock 3285109998Smarkm suffices. 3286109998Smarkm [Bodo Moeller] 3287109998Smarkm 3288109998Smarkm *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This 3289109998Smarkm sets the subject name for a new request or supersedes the 3290109998Smarkm subject name in a given request. Formats that can be parsed are 3291109998Smarkm 'CN=Some Name, OU=myOU, C=IT' 3292109998Smarkm and 3293109998Smarkm 'CN=Some Name/OU=myOU/C=IT'. 3294109998Smarkm 3295109998Smarkm Add options '-batch' and '-verbose' to 'openssl req'. 3296109998Smarkm [Massimiliano Pala <madwolf@hackmasters.net>] 3297109998Smarkm 3298109998Smarkm *) Introduce the possibility to access global variables through 3299109998Smarkm functions on platform were that's the best way to handle exporting 3300109998Smarkm global variables in shared libraries. To enable this functionality, 3301109998Smarkm one must configure with "EXPORT_VAR_AS_FN" or defined the C macro 3302109998Smarkm "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter 3303109998Smarkm is normally done by Configure or something similar). 3304109998Smarkm 3305109998Smarkm To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL 3306109998Smarkm in the source file (foo.c) like this: 3307109998Smarkm 3308109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1; 3309109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(double,bar); 3310109998Smarkm 3311109998Smarkm To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL 3312109998Smarkm and OPENSSL_GLOBAL_REF in the header file (foo.h) like this: 3313109998Smarkm 3314109998Smarkm OPENSSL_DECLARE_GLOBAL(int,foo); 3315109998Smarkm #define foo OPENSSL_GLOBAL_REF(foo) 3316109998Smarkm OPENSSL_DECLARE_GLOBAL(double,bar); 3317109998Smarkm #define bar OPENSSL_GLOBAL_REF(bar) 3318109998Smarkm 3319109998Smarkm The #defines are very important, and therefore so is including the 3320109998Smarkm header file everywhere where the defined globals are used. 3321109998Smarkm 3322109998Smarkm The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition 3323109998Smarkm of ASN.1 items, but that structure is a bit different. 3324109998Smarkm 3325109998Smarkm The largest change is in util/mkdef.pl which has been enhanced with 3326109998Smarkm better and easier to understand logic to choose which symbols should 3327109998Smarkm go into the Windows .def files as well as a number of fixes and code 3328109998Smarkm cleanup (among others, algorithm keywords are now sorted 3329109998Smarkm lexicographically to avoid constant rewrites). 3330109998Smarkm [Richard Levitte] 3331109998Smarkm 3332109998Smarkm *) In BN_div() keep a copy of the sign of 'num' before writing the 3333109998Smarkm result to 'rm' because if rm==num the value will be overwritten 3334109998Smarkm and produce the wrong result if 'num' is negative: this caused 3335109998Smarkm problems with BN_mod() and BN_nnmod(). 3336109998Smarkm [Steve Henson] 3337109998Smarkm 3338109998Smarkm *) Function OCSP_request_verify(). This checks the signature on an 3339109998Smarkm OCSP request and verifies the signer certificate. The signer 3340109998Smarkm certificate is just checked for a generic purpose and OCSP request 3341109998Smarkm trust settings. 3342109998Smarkm [Steve Henson] 3343109998Smarkm 3344109998Smarkm *) Add OCSP_check_validity() function to check the validity of OCSP 3345109998Smarkm responses. OCSP responses are prepared in real time and may only 3346109998Smarkm be a few seconds old. Simply checking that the current time lies 3347109998Smarkm between thisUpdate and nextUpdate max reject otherwise valid responses 3348109998Smarkm caused by either OCSP responder or client clock inaccuracy. Instead 3349109998Smarkm we allow thisUpdate and nextUpdate to fall within a certain period of 3350109998Smarkm the current time. The age of the response can also optionally be 3351109998Smarkm checked. Two new options -validity_period and -status_age added to 3352109998Smarkm ocsp utility. 3353109998Smarkm [Steve Henson] 3354109998Smarkm 3355109998Smarkm *) If signature or public key algorithm is unrecognized print out its 3356109998Smarkm OID rather that just UNKNOWN. 3357109998Smarkm [Steve Henson] 3358109998Smarkm 3359109998Smarkm *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and 3360109998Smarkm OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate 3361109998Smarkm ID to be generated from the issuer certificate alone which can then be 3362109998Smarkm passed to OCSP_id_issuer_cmp(). 3363109998Smarkm [Steve Henson] 3364109998Smarkm 3365109998Smarkm *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new 3366109998Smarkm ASN1 modules to export functions returning ASN1_ITEM pointers 3367109998Smarkm instead of the ASN1_ITEM structures themselves. This adds several 3368109998Smarkm new macros which allow the underlying ASN1 function/structure to 3369109998Smarkm be accessed transparently. As a result code should not use ASN1_ITEM 3370109998Smarkm references directly (such as &X509_it) but instead use the relevant 3371109998Smarkm macros (such as ASN1_ITEM_rptr(X509)). This option is to allow 3372109998Smarkm use of the new ASN1 code on platforms where exporting structures 3373109998Smarkm is problematical (for example in shared libraries) but exporting 3374109998Smarkm functions returning pointers to structures is not. 3375109998Smarkm [Steve Henson] 3376109998Smarkm 3377109998Smarkm *) Add support for overriding the generation of SSL/TLS session IDs. 3378109998Smarkm These callbacks can be registered either in an SSL_CTX or per SSL. 3379109998Smarkm The purpose of this is to allow applications to control, if they wish, 3380109998Smarkm the arbitrary values chosen for use as session IDs, particularly as it 3381109998Smarkm can be useful for session caching in multiple-server environments. A 3382109998Smarkm command-line switch for testing this (and any client code that wishes 3383109998Smarkm to use such a feature) has been added to "s_server". 3384109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 3385109998Smarkm 3386109998Smarkm *) Modify mkdef.pl to recognise and parse preprocessor conditionals 3387109998Smarkm of the form '#if defined(...) || defined(...) || ...' and 3388109998Smarkm '#if !defined(...) && !defined(...) && ...'. This also avoids 3389109998Smarkm the growing number of special cases it was previously handling. 3390109998Smarkm [Richard Levitte] 3391109998Smarkm 3392109998Smarkm *) Make all configuration macros available for application by making 3393109998Smarkm sure they are available in opensslconf.h, by giving them names starting 3394109998Smarkm with "OPENSSL_" to avoid conflicts with other packages and by making 3395109998Smarkm sure e_os2.h will cover all platform-specific cases together with 3396109998Smarkm opensslconf.h. 3397109998Smarkm Additionally, it is now possible to define configuration/platform- 3398109998Smarkm specific names (called "system identities"). In the C code, these 3399109998Smarkm are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another 3400109998Smarkm macro with the name beginning with "OPENSSL_SYS_", which is determined 3401109998Smarkm from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on 3402109998Smarkm what is available. 3403109998Smarkm [Richard Levitte] 3404109998Smarkm 3405109998Smarkm *) New option -set_serial to 'req' and 'x509' this allows the serial 3406109998Smarkm number to use to be specified on the command line. Previously self 3407109998Smarkm signed certificates were hard coded with serial number 0 and the 3408109998Smarkm CA options of 'x509' had to use a serial number in a file which was 3409109998Smarkm auto incremented. 3410109998Smarkm [Steve Henson] 3411109998Smarkm 3412109998Smarkm *) New options to 'ca' utility to support V2 CRL entry extensions. 3413109998Smarkm Currently CRL reason, invalidity date and hold instruction are 3414109998Smarkm supported. Add new CRL extensions to V3 code and some new objects. 3415109998Smarkm [Steve Henson] 3416109998Smarkm 3417109998Smarkm *) New function EVP_CIPHER_CTX_set_padding() this is used to 3418109998Smarkm disable standard block padding (aka PKCS#5 padding) in the EVP 3419109998Smarkm API, which was previously mandatory. This means that the data is 3420109998Smarkm not padded in any way and so the total length much be a multiple 3421109998Smarkm of the block size, otherwise an error occurs. 3422109998Smarkm [Steve Henson] 3423109998Smarkm 3424109998Smarkm *) Initial (incomplete) OCSP SSL support. 3425109998Smarkm [Steve Henson] 3426109998Smarkm 3427109998Smarkm *) New function OCSP_parse_url(). This splits up a URL into its host, 3428109998Smarkm port and path components: primarily to parse OCSP URLs. New -url 3429109998Smarkm option to ocsp utility. 3430109998Smarkm [Steve Henson] 3431109998Smarkm 3432109998Smarkm *) New nonce behavior. The return value of OCSP_check_nonce() now 3433109998Smarkm reflects the various checks performed. Applications can decide 3434109998Smarkm whether to tolerate certain situations such as an absent nonce 3435109998Smarkm in a response when one was present in a request: the ocsp application 3436109998Smarkm just prints out a warning. New function OCSP_add1_basic_nonce() 3437109998Smarkm this is to allow responders to include a nonce in a response even if 3438109998Smarkm the request is nonce-less. 3439109998Smarkm [Steve Henson] 3440109998Smarkm 3441109998Smarkm *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are 3442109998Smarkm skipped when using openssl x509 multiple times on a single input file, 3443109998Smarkm e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs". 3444109998Smarkm [Bodo Moeller] 3445109998Smarkm 3446109998Smarkm *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string() 3447109998Smarkm set string type: to handle setting ASN1_TIME structures. Fix ca 3448109998Smarkm utility to correctly initialize revocation date of CRLs. 3449109998Smarkm [Steve Henson] 3450109998Smarkm 3451109998Smarkm *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override 3452109998Smarkm the clients preferred ciphersuites and rather use its own preferences. 3453109998Smarkm Should help to work around M$ SGC (Server Gated Cryptography) bug in 3454109998Smarkm Internet Explorer by ensuring unchanged hash method during stepup. 3455109998Smarkm (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.) 3456109998Smarkm [Lutz Jaenicke] 3457109998Smarkm 3458109998Smarkm *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael 3459109998Smarkm to aes and add a new 'exist' option to print out symbols that don't 3460109998Smarkm appear to exist. 3461109998Smarkm [Steve Henson] 3462109998Smarkm 3463109998Smarkm *) Additional options to ocsp utility to allow flags to be set and 3464109998Smarkm additional certificates supplied. 3465109998Smarkm [Steve Henson] 3466109998Smarkm 3467109998Smarkm *) Add the option -VAfile to 'openssl ocsp', so the user can give the 3468109998Smarkm OCSP client a number of certificate to only verify the response 3469109998Smarkm signature against. 3470109998Smarkm [Richard Levitte] 3471109998Smarkm 3472109998Smarkm *) Update Rijndael code to version 3.0 and change EVP AES ciphers to 3473109998Smarkm handle the new API. Currently only ECB, CBC modes supported. Add new 3474109998Smarkm AES OIDs. 3475109998Smarkm 3476109998Smarkm Add TLS AES ciphersuites as described in RFC3268, "Advanced 3477109998Smarkm Encryption Standard (AES) Ciphersuites for Transport Layer 3478109998Smarkm Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were 3479109998Smarkm not enabled by default and were not part of the "ALL" ciphersuite 3480109998Smarkm alias because they were not yet official; they could be 3481109998Smarkm explicitly requested by specifying the "AESdraft" ciphersuite 3482109998Smarkm group alias. In the final release of OpenSSL 0.9.7, the group 3483109998Smarkm alias is called "AES" and is part of "ALL".) 3484109998Smarkm [Ben Laurie, Steve Henson, Bodo Moeller] 3485109998Smarkm 3486109998Smarkm *) New function OCSP_copy_nonce() to copy nonce value (if present) from 3487109998Smarkm request to response. 3488109998Smarkm [Steve Henson] 3489109998Smarkm 3490109998Smarkm *) Functions for OCSP responders. OCSP_request_onereq_count(), 3491109998Smarkm OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info() 3492109998Smarkm extract information from a certificate request. OCSP_response_create() 3493109998Smarkm creates a response and optionally adds a basic response structure. 3494109998Smarkm OCSP_basic_add1_status() adds a complete single response to a basic 3495109998Smarkm response and returns the OCSP_SINGLERESP structure just added (to allow 3496109998Smarkm extensions to be included for example). OCSP_basic_add1_cert() adds a 3497109998Smarkm certificate to a basic response and OCSP_basic_sign() signs a basic 3498109998Smarkm response with various flags. New helper functions ASN1_TIME_check() 3499109998Smarkm (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime() 3500109998Smarkm (converts ASN1_TIME to GeneralizedTime). 3501109998Smarkm [Steve Henson] 3502109998Smarkm 3503109998Smarkm *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}() 3504109998Smarkm in a single operation. X509_get0_pubkey_bitstr() extracts the public_key 3505109998Smarkm structure from a certificate. X509_pubkey_digest() digests the public_key 3506109998Smarkm contents: this is used in various key identifiers. 3507109998Smarkm [Steve Henson] 3508109998Smarkm 3509109998Smarkm *) Make sk_sort() tolerate a NULL argument. 3510109998Smarkm [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>] 3511109998Smarkm 3512109998Smarkm *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates 3513109998Smarkm passed by the function are trusted implicitly. If any of them signed the 3514109998Smarkm response then it is assumed to be valid and is not verified. 3515109998Smarkm [Steve Henson] 3516109998Smarkm 3517109998Smarkm *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT 3518109998Smarkm to data. This was previously part of the PKCS7 ASN1 code. This 3519109998Smarkm was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures. 3520109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 3521109998Smarkm <support@securenetterm.com>] 3522109998Smarkm 3523109998Smarkm *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1 3524109998Smarkm routines: without these tracing memory leaks is very painful. 3525109998Smarkm Fix leaks in PKCS12 and PKCS7 routines. 3526109998Smarkm [Steve Henson] 3527109998Smarkm 3528109998Smarkm *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new(). 3529109998Smarkm Previously it initialised the 'type' argument to V_ASN1_UTCTIME which 3530109998Smarkm effectively meant GeneralizedTime would never be used. Now it 3531109998Smarkm is initialised to -1 but X509_time_adj() now has to check the value 3532109998Smarkm and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or 3533109998Smarkm V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime. 3534109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 3535109998Smarkm <support@securenetterm.com>] 3536109998Smarkm 3537109998Smarkm *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously 3538109998Smarkm result in a zero length in the ASN1_INTEGER structure which was 3539109998Smarkm not consistent with the structure when d2i_ASN1_INTEGER() was used 3540109998Smarkm and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER() 3541109998Smarkm to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER() 3542109998Smarkm where it did not print out a minus for negative ASN1_INTEGER. 3543109998Smarkm [Steve Henson] 3544109998Smarkm 3545109998Smarkm *) Add summary printout to ocsp utility. The various functions which 3546109998Smarkm convert status values to strings have been renamed to: 3547109998Smarkm OCSP_response_status_str(), OCSP_cert_status_str() and 3548109998Smarkm OCSP_crl_reason_str() and are no longer static. New options 3549109998Smarkm to verify nonce values and to disable verification. OCSP response 3550109998Smarkm printout format cleaned up. 3551109998Smarkm [Steve Henson] 3552109998Smarkm 3553109998Smarkm *) Add additional OCSP certificate checks. These are those specified 3554109998Smarkm in RFC2560. This consists of two separate checks: the CA of the 3555109998Smarkm certificate being checked must either be the OCSP signer certificate 3556109998Smarkm or the issuer of the OCSP signer certificate. In the latter case the 3557109998Smarkm OCSP signer certificate must contain the OCSP signing extended key 3558109998Smarkm usage. This check is performed by attempting to match the OCSP 3559109998Smarkm signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash 3560109998Smarkm in the OCSP_CERTID structures of the response. 3561109998Smarkm [Steve Henson] 3562109998Smarkm 3563109998Smarkm *) Initial OCSP certificate verification added to OCSP_basic_verify() 3564109998Smarkm and related routines. This uses the standard OpenSSL certificate 3565109998Smarkm verify routines to perform initial checks (just CA validity) and 3566109998Smarkm to obtain the certificate chain. Then additional checks will be 3567109998Smarkm performed on the chain. Currently the root CA is checked to see 3568109998Smarkm if it is explicitly trusted for OCSP signing. This is used to set 3569109998Smarkm a root CA as a global signing root: that is any certificate that 3570109998Smarkm chains to that CA is an acceptable OCSP signing certificate. 3571109998Smarkm [Steve Henson] 3572109998Smarkm 3573109998Smarkm *) New '-extfile ...' option to 'openssl ca' for reading X.509v3 3574109998Smarkm extensions from a separate configuration file. 3575109998Smarkm As when reading extensions from the main configuration file, 3576109998Smarkm the '-extensions ...' option may be used for specifying the 3577109998Smarkm section to use. 3578109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 3579109998Smarkm 3580109998Smarkm *) New OCSP utility. Allows OCSP requests to be generated or 3581109998Smarkm read. The request can be sent to a responder and the output 3582109998Smarkm parsed, outputed or printed in text form. Not complete yet: 3583109998Smarkm still needs to check the OCSP response validity. 3584109998Smarkm [Steve Henson] 3585109998Smarkm 3586109998Smarkm *) New subcommands for 'openssl ca': 3587109998Smarkm 'openssl ca -status <serial>' prints the status of the cert with 3588109998Smarkm the given serial number (according to the index file). 3589109998Smarkm 'openssl ca -updatedb' updates the expiry status of certificates 3590109998Smarkm in the index file. 3591109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 3592109998Smarkm 3593109998Smarkm *) New '-newreq-nodes' command option to CA.pl. This is like 3594109998Smarkm '-newreq', but calls 'openssl req' with the '-nodes' option 3595109998Smarkm so that the resulting key is not encrypted. 3596109998Smarkm [Damien Miller <djm@mindrot.org>] 3597109998Smarkm 3598109998Smarkm *) New configuration for the GNU Hurd. 3599109998Smarkm [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte] 3600109998Smarkm 3601109998Smarkm *) Initial code to implement OCSP basic response verify. This 3602109998Smarkm is currently incomplete. Currently just finds the signer's 3603109998Smarkm certificate and verifies the signature on the response. 3604109998Smarkm [Steve Henson] 3605109998Smarkm 3606109998Smarkm *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in 3607109998Smarkm value of OPENSSLDIR. This is available via the new '-d' option 3608109998Smarkm to 'openssl version', and is also included in 'openssl version -a'. 3609109998Smarkm [Bodo Moeller] 3610109998Smarkm 3611109998Smarkm *) Allowing defining memory allocation callbacks that will be given 3612109998Smarkm file name and line number information in additional arguments 3613109998Smarkm (a const char* and an int). The basic functionality remains, as 3614109998Smarkm well as the original possibility to just replace malloc(), 3615109998Smarkm realloc() and free() by functions that do not know about these 3616109998Smarkm additional arguments. To register and find out the current 3617109998Smarkm settings for extended allocation functions, the following 3618109998Smarkm functions are provided: 3619109998Smarkm 3620109998Smarkm CRYPTO_set_mem_ex_functions 3621109998Smarkm CRYPTO_set_locked_mem_ex_functions 3622109998Smarkm CRYPTO_get_mem_ex_functions 3623109998Smarkm CRYPTO_get_locked_mem_ex_functions 3624109998Smarkm 3625109998Smarkm These work the same way as CRYPTO_set_mem_functions and friends. 3626109998Smarkm CRYPTO_get_[locked_]mem_functions now writes 0 where such an 3627109998Smarkm extended allocation function is enabled. 3628109998Smarkm Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where 3629109998Smarkm a conventional allocation function is enabled. 3630109998Smarkm [Richard Levitte, Bodo Moeller] 3631109998Smarkm 3632109998Smarkm *) Finish off removing the remaining LHASH function pointer casts. 3633109998Smarkm There should no longer be any prototype-casting required when using 3634109998Smarkm the LHASH abstraction, and any casts that remain are "bugs". See 3635109998Smarkm the callback types and macros at the head of lhash.h for details 3636109998Smarkm (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example). 3637109998Smarkm [Geoff Thorpe] 3638109998Smarkm 3639109998Smarkm *) Add automatic query of EGD sockets in RAND_poll() for the unix variant. 3640109998Smarkm If /dev/[u]random devices are not available or do not return enough 3641109998Smarkm entropy, EGD style sockets (served by EGD or PRNGD) will automatically 3642109998Smarkm be queried. 3643109998Smarkm The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and 3644109998Smarkm /etc/entropy will be queried once each in this sequence, quering stops 3645109998Smarkm when enough entropy was collected without querying more sockets. 3646109998Smarkm [Lutz Jaenicke] 3647109998Smarkm 3648109998Smarkm *) Change the Unix RAND_poll() variant to be able to poll several 3649109998Smarkm random devices, as specified by DEVRANDOM, until a sufficient amount 3650109998Smarkm of data has been collected. We spend at most 10 ms on each file 3651109998Smarkm (select timeout) and read in non-blocking mode. DEVRANDOM now 3652109998Smarkm defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom" 3653109998Smarkm (previously it was just the string "/dev/urandom"), so on typical 3654109998Smarkm platforms the 10 ms delay will never occur. 3655109998Smarkm Also separate out the Unix variant to its own file, rand_unix.c. 3656109998Smarkm For VMS, there's a currently-empty rand_vms.c. 3657109998Smarkm [Richard Levitte] 3658109998Smarkm 3659109998Smarkm *) Move OCSP client related routines to ocsp_cl.c. These 3660109998Smarkm provide utility functions which an application needing 3661109998Smarkm to issue a request to an OCSP responder and analyse the 3662109998Smarkm response will typically need: as opposed to those which an 3663109998Smarkm OCSP responder itself would need which will be added later. 3664109998Smarkm 3665109998Smarkm OCSP_request_sign() signs an OCSP request with an API similar 3666109998Smarkm to PKCS7_sign(). OCSP_response_status() returns status of OCSP 3667109998Smarkm response. OCSP_response_get1_basic() extracts basic response 3668109998Smarkm from response. OCSP_resp_find_status(): finds and extracts status 3669109998Smarkm information from an OCSP_CERTID structure (which will be created 3670109998Smarkm when the request structure is built). These are built from lower 3671109998Smarkm level functions which work on OCSP_SINGLERESP structures but 3672109998Smarkm wont normally be used unless the application wishes to examine 3673109998Smarkm extensions in the OCSP response for example. 3674109998Smarkm 3675109998Smarkm Replace nonce routines with a pair of functions. 3676109998Smarkm OCSP_request_add1_nonce() adds a nonce value and optionally 3677109998Smarkm generates a random value. OCSP_check_nonce() checks the 3678109998Smarkm validity of the nonce in an OCSP response. 3679109998Smarkm [Steve Henson] 3680109998Smarkm 3681109998Smarkm *) Change function OCSP_request_add() to OCSP_request_add0_id(). 3682109998Smarkm This doesn't copy the supplied OCSP_CERTID and avoids the 3683109998Smarkm need to free up the newly created id. Change return type 3684109998Smarkm to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure. 3685109998Smarkm This can then be used to add extensions to the request. 3686109998Smarkm Deleted OCSP_request_new(), since most of its functionality 3687109998Smarkm is now in OCSP_REQUEST_new() (and the case insensitive name 3688109998Smarkm clash) apart from the ability to set the request name which 3689109998Smarkm will be added elsewhere. 3690109998Smarkm [Steve Henson] 3691109998Smarkm 3692109998Smarkm *) Update OCSP API. Remove obsolete extensions argument from 3693109998Smarkm various functions. Extensions are now handled using the new 3694109998Smarkm OCSP extension code. New simple OCSP HTTP function which 3695109998Smarkm can be used to send requests and parse the response. 3696109998Smarkm [Steve Henson] 3697109998Smarkm 3698109998Smarkm *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new 3699109998Smarkm ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN 3700109998Smarkm uses the special reorder version of SET OF to sort the attributes 3701109998Smarkm and reorder them to match the encoded order. This resolves a long 3702109998Smarkm standing problem: a verify on a PKCS7 structure just after signing 3703109998Smarkm it used to fail because the attribute order did not match the 3704109998Smarkm encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes: 3705109998Smarkm it uses the received order. This is necessary to tolerate some broken 3706109998Smarkm software that does not order SET OF. This is handled by encoding 3707109998Smarkm as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class) 3708109998Smarkm to produce the required SET OF. 3709109998Smarkm [Steve Henson] 3710109998Smarkm 3711109998Smarkm *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and 3712109998Smarkm OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header 3713109998Smarkm files to get correct declarations of the ASN.1 item variables. 3714109998Smarkm [Richard Levitte] 3715109998Smarkm 3716109998Smarkm *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many 3717109998Smarkm PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs: 3718109998Smarkm asn1_check_tlen() would sometimes attempt to use 'ctx' when it was 3719109998Smarkm NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i(). 3720109998Smarkm New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant 3721109998Smarkm ASN1_ITEM and no wrapper functions. 3722109998Smarkm [Steve Henson] 3723109998Smarkm 3724109998Smarkm *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These 3725109998Smarkm replace the old function pointer based I/O routines. Change most of 3726109998Smarkm the *_d2i_bio() and *_d2i_fp() functions to use these. 3727109998Smarkm [Steve Henson] 3728109998Smarkm 3729109998Smarkm *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor 3730109998Smarkm lines, recognice more "algorithms" that can be deselected, and make 3731109998Smarkm it complain about algorithm deselection that isn't recognised. 3732109998Smarkm [Richard Levitte] 3733109998Smarkm 3734109998Smarkm *) New ASN1 functions to handle dup, sign, verify, digest, pack and 3735109998Smarkm unpack operations in terms of ASN1_ITEM. Modify existing wrappers 3736109998Smarkm to use new functions. Add NO_ASN1_OLD which can be set to remove 3737109998Smarkm some old style ASN1 functions: this can be used to determine if old 3738109998Smarkm code will still work when these eventually go away. 3739109998Smarkm [Steve Henson] 3740109998Smarkm 3741109998Smarkm *) New extension functions for OCSP structures, these follow the 3742109998Smarkm same conventions as certificates and CRLs. 3743109998Smarkm [Steve Henson] 3744109998Smarkm 3745109998Smarkm *) New function X509V3_add1_i2d(). This automatically encodes and 3746109998Smarkm adds an extension. Its behaviour can be customised with various 3747109998Smarkm flags to append, replace or delete. Various wrappers added for 3748109998Smarkm certifcates and CRLs. 3749109998Smarkm [Steve Henson] 3750109998Smarkm 3751109998Smarkm *) Fix to avoid calling the underlying ASN1 print routine when 3752109998Smarkm an extension cannot be parsed. Correct a typo in the 3753109998Smarkm OCSP_SERVICELOC extension. Tidy up print OCSP format. 3754109998Smarkm [Steve Henson] 3755109998Smarkm 3756109998Smarkm *) Make mkdef.pl parse some of the ASN1 macros and add apropriate 3757109998Smarkm entries for variables. 3758109998Smarkm [Steve Henson] 3759109998Smarkm 3760109998Smarkm *) Add functionality to apps/openssl.c for detecting locking 3761109998Smarkm problems: As the program is single-threaded, all we have 3762109998Smarkm to do is register a locking callback using an array for 3763109998Smarkm storing which locks are currently held by the program. 3764109998Smarkm [Bodo Moeller] 3765109998Smarkm 3766109998Smarkm *) Use a lock around the call to CRYPTO_get_ex_new_index() in 3767109998Smarkm SSL_get_ex_data_X509_STORE_idx(), which is used in 3768109998Smarkm ssl_verify_cert_chain() and thus can be called at any time 3769109998Smarkm during TLS/SSL handshakes so that thread-safety is essential. 3770109998Smarkm Unfortunately, the ex_data design is not at all suited 3771109998Smarkm for multi-threaded use, so it probably should be abolished. 3772109998Smarkm [Bodo Moeller] 3773109998Smarkm 3774109998Smarkm *) Added Broadcom "ubsec" ENGINE to OpenSSL. 3775109998Smarkm [Broadcom, tweaked and integrated by Geoff Thorpe] 3776109998Smarkm 3777109998Smarkm *) Move common extension printing code to new function 3778109998Smarkm X509V3_print_extensions(). Reorganise OCSP print routines and 3779109998Smarkm implement some needed OCSP ASN1 functions. Add OCSP extensions. 3780109998Smarkm [Steve Henson] 3781109998Smarkm 3782109998Smarkm *) New function X509_signature_print() to remove duplication in some 3783109998Smarkm print routines. 3784109998Smarkm [Steve Henson] 3785109998Smarkm 3786109998Smarkm *) Add a special meaning when SET OF and SEQUENCE OF flags are both 3787109998Smarkm set (this was treated exactly the same as SET OF previously). This 3788109998Smarkm is used to reorder the STACK representing the structure to match the 3789109998Smarkm encoding. This will be used to get round a problem where a PKCS7 3790109998Smarkm structure which was signed could not be verified because the STACK 3791109998Smarkm order did not reflect the encoded order. 3792109998Smarkm [Steve Henson] 3793109998Smarkm 3794109998Smarkm *) Reimplement the OCSP ASN1 module using the new code. 3795109998Smarkm [Steve Henson] 3796109998Smarkm 3797109998Smarkm *) Update the X509V3 code to permit the use of an ASN1_ITEM structure 3798109998Smarkm for its ASN1 operations. The old style function pointers still exist 3799109998Smarkm for now but they will eventually go away. 3800109998Smarkm [Steve Henson] 3801109998Smarkm 3802109998Smarkm *) Merge in replacement ASN1 code from the ASN1 branch. This almost 3803109998Smarkm completely replaces the old ASN1 functionality with a table driven 3804109998Smarkm encoder and decoder which interprets an ASN1_ITEM structure describing 3805109998Smarkm the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is 3806109998Smarkm largely maintained. Almost all of the old asn1_mac.h macro based ASN1 3807109998Smarkm has also been converted to the new form. 3808109998Smarkm [Steve Henson] 3809109998Smarkm 3810109998Smarkm *) Change BN_mod_exp_recp so that negative moduli are tolerated 3811109998Smarkm (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set 3812109998Smarkm so that BN_mod_exp_mont and BN_mod_exp_mont_word work 3813109998Smarkm for negative moduli. 3814109998Smarkm [Bodo Moeller] 3815109998Smarkm 3816109998Smarkm *) Fix BN_uadd and BN_usub: Always return non-negative results instead 3817109998Smarkm of not touching the result's sign bit. 3818109998Smarkm [Bodo Moeller] 3819109998Smarkm 3820109998Smarkm *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be 3821109998Smarkm set. 3822109998Smarkm [Bodo Moeller] 3823109998Smarkm 3824109998Smarkm *) Changed the LHASH code to use prototypes for callbacks, and created 3825109998Smarkm macros to declare and implement thin (optionally static) functions 3826109998Smarkm that provide type-safety and avoid function pointer casting for the 3827109998Smarkm type-specific callbacks. 3828109998Smarkm [Geoff Thorpe] 3829109998Smarkm 3830109998Smarkm *) Added Kerberos Cipher Suites to be used with TLS, as written in 3831109998Smarkm RFC 2712. 3832109998Smarkm [Veers Staats <staatsvr@asc.hpc.mil>, 3833109998Smarkm Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte] 3834109998Smarkm 3835109998Smarkm *) Reformat the FAQ so the different questions and answers can be divided 3836109998Smarkm in sections depending on the subject. 3837109998Smarkm [Richard Levitte] 3838109998Smarkm 3839109998Smarkm *) Have the zlib compression code load ZLIB.DLL dynamically under 3840109998Smarkm Windows. 3841109998Smarkm [Richard Levitte] 3842109998Smarkm 3843109998Smarkm *) New function BN_mod_sqrt for computing square roots modulo a prime 3844109998Smarkm (using the probabilistic Tonelli-Shanks algorithm unless 3845109998Smarkm p == 3 (mod 4) or p == 5 (mod 8), which are cases that can 3846109998Smarkm be handled deterministically). 3847109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 3848109998Smarkm 3849109998Smarkm *) Make BN_mod_inverse faster by explicitly handling small quotients 3850109998Smarkm in the Euclid loop. (Speed gain about 20% for small moduli [256 or 3851109998Smarkm 512 bits], about 30% for larger ones [1024 or 2048 bits].) 3852109998Smarkm [Bodo Moeller] 3853109998Smarkm 3854109998Smarkm *) New function BN_kronecker. 3855109998Smarkm [Bodo Moeller] 3856109998Smarkm 3857109998Smarkm *) Fix BN_gcd so that it works on negative inputs; the result is 3858109998Smarkm positive unless both parameters are zero. 3859109998Smarkm Previously something reasonably close to an infinite loop was 3860109998Smarkm possible because numbers could be growing instead of shrinking 3861109998Smarkm in the implementation of Euclid's algorithm. 3862109998Smarkm [Bodo Moeller] 3863109998Smarkm 3864109998Smarkm *) Fix BN_is_word() and BN_is_one() macros to take into account the 3865109998Smarkm sign of the number in question. 3866109998Smarkm 3867109998Smarkm Fix BN_is_word(a,w) to work correctly for w == 0. 3868109998Smarkm 3869109998Smarkm The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w) 3870109998Smarkm because its test if the absolute value of 'a' equals 'w'. 3871109998Smarkm Note that BN_abs_is_word does *not* handle w == 0 reliably; 3872109998Smarkm it exists mostly for use in the implementations of BN_is_zero(), 3873109998Smarkm BN_is_one(), and BN_is_word(). 3874109998Smarkm [Bodo Moeller] 3875109998Smarkm 3876109998Smarkm *) New function BN_swap. 3877109998Smarkm [Bodo Moeller] 3878109998Smarkm 3879109998Smarkm *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that 3880109998Smarkm the exponentiation functions are more likely to produce reasonable 3881109998Smarkm results on negative inputs. 3882109998Smarkm [Bodo Moeller] 3883109998Smarkm 3884109998Smarkm *) Change BN_mod_mul so that the result is always non-negative. 3885109998Smarkm Previously, it could be negative if one of the factors was negative; 3886109998Smarkm I don't think anyone really wanted that behaviour. 3887109998Smarkm [Bodo Moeller] 3888109998Smarkm 3889109998Smarkm *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c 3890109998Smarkm (except for exponentiation, which stays in crypto/bn/bn_exp.c, 3891109998Smarkm and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c) 3892109998Smarkm and add new functions: 3893109998Smarkm 3894109998Smarkm BN_nnmod 3895109998Smarkm BN_mod_sqr 3896109998Smarkm BN_mod_add 3897109998Smarkm BN_mod_add_quick 3898109998Smarkm BN_mod_sub 3899109998Smarkm BN_mod_sub_quick 3900109998Smarkm BN_mod_lshift1 3901109998Smarkm BN_mod_lshift1_quick 3902109998Smarkm BN_mod_lshift 3903109998Smarkm BN_mod_lshift_quick 3904109998Smarkm 3905109998Smarkm These functions always generate non-negative results. 3906109998Smarkm 3907109998Smarkm BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r 3908109998Smarkm such that |m| < r < 0, BN_nnmod will output rem + |m| instead). 3909109998Smarkm 3910109998Smarkm BN_mod_XXX_quick(r, a, [b,] m) generates the same result as 3911109998Smarkm BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b] 3912109998Smarkm be reduced modulo m. 3913109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 3914109998Smarkm 3915111147Snectar#if 0 3916111147Snectar The following entry accidentily appeared in the CHANGES file 3917111147Snectar distributed with OpenSSL 0.9.7. The modifications described in 3918111147Snectar it do *not* apply to OpenSSL 0.9.7. 3919111147Snectar 3920109998Smarkm *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 3921109998Smarkm was actually never needed) and in BN_mul(). The removal in BN_mul() 3922109998Smarkm required a small change in bn_mul_part_recursive() and the addition 3923109998Smarkm of the functions bn_cmp_part_words(), bn_sub_part_words() and 3924109998Smarkm bn_add_part_words(), which do the same thing as bn_cmp_words(), 3925109998Smarkm bn_sub_words() and bn_add_words() except they take arrays with 3926109998Smarkm differing sizes. 3927109998Smarkm [Richard Levitte] 3928111147Snectar#endif 3929109998Smarkm 3930109998Smarkm *) In 'openssl passwd', verify passwords read from the terminal 3931109998Smarkm unless the '-salt' option is used (which usually means that 3932109998Smarkm verification would just waste user's time since the resulting 3933109998Smarkm hash is going to be compared with some given password hash) 3934109998Smarkm or the new '-noverify' option is used. 3935109998Smarkm 3936109998Smarkm This is an incompatible change, but it does not affect 3937109998Smarkm non-interactive use of 'openssl passwd' (passwords on the command 3938109998Smarkm line, '-stdin' option, '-in ...' option) and thus should not 3939109998Smarkm cause any problems. 3940109998Smarkm [Bodo Moeller] 3941109998Smarkm 3942109998Smarkm *) Remove all references to RSAref, since there's no more need for it. 3943109998Smarkm [Richard Levitte] 3944109998Smarkm 3945109998Smarkm *) Make DSO load along a path given through an environment variable 3946109998Smarkm (SHLIB_PATH) with shl_load(). 3947109998Smarkm [Richard Levitte] 3948109998Smarkm 3949109998Smarkm *) Constify the ENGINE code as a result of BIGNUM constification. 3950109998Smarkm Also constify the RSA code and most things related to it. In a 3951109998Smarkm few places, most notable in the depth of the ASN.1 code, ugly 3952109998Smarkm casts back to non-const were required (to be solved at a later 3953109998Smarkm time) 3954109998Smarkm [Richard Levitte] 3955109998Smarkm 3956109998Smarkm *) Make it so the openssl application has all engines loaded by default. 3957109998Smarkm [Richard Levitte] 3958109998Smarkm 3959109998Smarkm *) Constify the BIGNUM routines a little more. 3960109998Smarkm [Richard Levitte] 3961109998Smarkm 3962109998Smarkm *) Add the following functions: 3963109998Smarkm 3964109998Smarkm ENGINE_load_cswift() 3965109998Smarkm ENGINE_load_chil() 3966109998Smarkm ENGINE_load_atalla() 3967109998Smarkm ENGINE_load_nuron() 3968109998Smarkm ENGINE_load_builtin_engines() 3969109998Smarkm 3970109998Smarkm That way, an application can itself choose if external engines that 3971109998Smarkm are built-in in OpenSSL shall ever be used or not. The benefit is 3972109998Smarkm that applications won't have to be linked with libdl or other dso 3973109998Smarkm libraries unless it's really needed. 3974109998Smarkm 3975109998Smarkm Changed 'openssl engine' to load all engines on demand. 3976109998Smarkm Changed the engine header files to avoid the duplication of some 3977109998Smarkm declarations (they differed!). 3978109998Smarkm [Richard Levitte] 3979109998Smarkm 3980109998Smarkm *) 'openssl engine' can now list capabilities. 3981109998Smarkm [Richard Levitte] 3982109998Smarkm 3983109998Smarkm *) Better error reporting in 'openssl engine'. 3984109998Smarkm [Richard Levitte] 3985109998Smarkm 3986109998Smarkm *) Never call load_dh_param(NULL) in s_server. 3987109998Smarkm [Bodo Moeller] 3988109998Smarkm 3989109998Smarkm *) Add engine application. It can currently list engines by name and 3990109998Smarkm identity, and test if they are actually available. 3991109998Smarkm [Richard Levitte] 3992109998Smarkm 3993109998Smarkm *) Improve RPM specification file by forcing symbolic linking and making 3994109998Smarkm sure the installed documentation is also owned by root.root. 3995109998Smarkm [Damien Miller <djm@mindrot.org>] 3996109998Smarkm 3997109998Smarkm *) Give the OpenSSL applications more possibilities to make use of 3998109998Smarkm keys (public as well as private) handled by engines. 3999109998Smarkm [Richard Levitte] 4000109998Smarkm 4001109998Smarkm *) Add OCSP code that comes from CertCo. 4002109998Smarkm [Richard Levitte] 4003109998Smarkm 4004109998Smarkm *) Add VMS support for the Rijndael code. 4005109998Smarkm [Richard Levitte] 4006109998Smarkm 4007109998Smarkm *) Added untested support for Nuron crypto accelerator. 4008109998Smarkm [Ben Laurie] 4009109998Smarkm 4010109998Smarkm *) Add support for external cryptographic devices. This code was 4011109998Smarkm previously distributed separately as the "engine" branch. 4012109998Smarkm [Geoff Thorpe, Richard Levitte] 4013109998Smarkm 4014109998Smarkm *) Rework the filename-translation in the DSO code. It is now possible to 4015109998Smarkm have far greater control over how a "name" is turned into a filename 4016109998Smarkm depending on the operating environment and any oddities about the 4017109998Smarkm different shared library filenames on each system. 4018109998Smarkm [Geoff Thorpe] 4019109998Smarkm 4020109998Smarkm *) Support threads on FreeBSD-elf in Configure. 4021109998Smarkm [Richard Levitte] 4022109998Smarkm 4023109998Smarkm *) Fix for SHA1 assembly problem with MASM: it produces 4024109998Smarkm warnings about corrupt line number information when assembling 4025109998Smarkm with debugging information. This is caused by the overlapping 4026109998Smarkm of two sections. 4027109998Smarkm [Bernd Matthes <mainbug@celocom.de>, Steve Henson] 4028109998Smarkm 4029109998Smarkm *) NCONF changes. 4030109998Smarkm NCONF_get_number() has no error checking at all. As a replacement, 4031109998Smarkm NCONF_get_number_e() is defined (_e for "error checking") and is 4032109998Smarkm promoted strongly. The old NCONF_get_number is kept around for 4033109998Smarkm binary backward compatibility. 4034109998Smarkm Make it possible for methods to load from something other than a BIO, 4035109998Smarkm by providing a function pointer that is given a name instead of a BIO. 4036109998Smarkm For example, this could be used to load configuration data from an 4037109998Smarkm LDAP server. 4038109998Smarkm [Richard Levitte] 4039109998Smarkm 4040109998Smarkm *) Fix for non blocking accept BIOs. Added new I/O special reason 4041109998Smarkm BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs 4042109998Smarkm with non blocking I/O was not possible because no retry code was 4043109998Smarkm implemented. Also added new SSL code SSL_WANT_ACCEPT to cover 4044109998Smarkm this case. 4045109998Smarkm [Steve Henson] 4046109998Smarkm 4047109998Smarkm *) Added the beginnings of Rijndael support. 4048109998Smarkm [Ben Laurie] 4049109998Smarkm 4050109998Smarkm *) Fix for bug in DirectoryString mask setting. Add support for 4051109998Smarkm X509_NAME_print_ex() in 'req' and X509_print_ex() function 4052109998Smarkm to allow certificate printing to more controllable, additional 4053109998Smarkm 'certopt' option to 'x509' to allow new printing options to be 4054109998Smarkm set. 4055109998Smarkm [Steve Henson] 4056109998Smarkm 4057109998Smarkm *) Clean old EAY MD5 hack from e_os.h. 4058109998Smarkm [Richard Levitte] 4059109998Smarkm 4060142425Snectar Changes between 0.9.6l and 0.9.6m [17 Mar 2004] 4061142425Snectar 4062142425Snectar *) Fix null-pointer assignment in do_change_cipher_spec() revealed 4063160814Ssimon by using the Codenomicon TLS Test Tool (CVE-2004-0079) 4064142425Snectar [Joe Orton, Steve Henson] 4065142425Snectar 4066142425Snectar Changes between 0.9.6k and 0.9.6l [04 Nov 2003] 4067142425Snectar 4068142425Snectar *) Fix additional bug revealed by the NISCC test suite: 4069142425Snectar 4070142425Snectar Stop bug triggering large recursion when presented with 4071160814Ssimon certain ASN.1 tags (CVE-2003-0851) 4072142425Snectar [Steve Henson] 4073142425Snectar 4074120631Snectar Changes between 0.9.6j and 0.9.6k [30 Sep 2003] 4075120631Snectar 4076120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 4077120631Snectar 4078120631Snectar Stop out of bounds reads in the ASN1 code when presented with 4079160814Ssimon invalid tags (CVE-2003-0543 and CVE-2003-0544). 4080120631Snectar 4081120631Snectar If verify callback ignores invalid public key errors don't try to check 4082120631Snectar certificate signature with the NULL public key. 4083120631Snectar 4084120631Snectar [Steve Henson] 4085120631Snectar 4086120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 4087120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 4088120631Snectar specifications. 4089120631Snectar [Steve Henson] 4090120631Snectar 4091120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 4092120631Snectar extra data after the compression methods not only for TLS 1.0 4093120631Snectar but also for SSL 3.0 (as required by the specification). 4094120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 4095120631Snectar 4096120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 4097120631Snectar when it's 512 *bits* long, not 512 bytes. 4098120631Snectar [Richard Levitte] 4099120631Snectar 4100120631Snectar Changes between 0.9.6i and 0.9.6j [10 Apr 2003] 4101120631Snectar 4102120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 4103120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 4104120631Snectar a protocol version number mismatch like a decryption error 4105120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 4106120631Snectar [Bodo Moeller] 4107120631Snectar 4108120631Snectar *) Turn on RSA blinding by default in the default implementation 4109120631Snectar to avoid a timing attack. Applications that don't want it can call 4110120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 4111120631Snectar They would be ill-advised to do so in most cases. 4112120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 4113120631Snectar 4114120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 4115120631Snectar seeded (in this case, the secret RSA exponent is abused as 4116120631Snectar an unpredictable seed -- if it is not unpredictable, there 4117120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 4118120631Snectar by remembering the creator's thread ID in rsa->blinding and 4119120631Snectar having all other threads use local one-time blinding factors 4120120631Snectar (this requires more computation than sharing rsa->blinding, but 4121120631Snectar avoids excessive locking; and if an RSA object is not shared 4122120631Snectar between threads, blinding will still be very fast). 4123120631Snectar [Bodo Moeller] 4124120631Snectar 4125111147Snectar Changes between 0.9.6h and 0.9.6i [19 Feb 2003] 4126111147Snectar 4127111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 4128111147Snectar via timing by performing a MAC computation even if incorrrect 4129111147Snectar block cipher padding has been found. This is a countermeasure 4130111147Snectar against active attacks where the attacker has to distinguish 4131160814Ssimon between bad padding and a MAC verification error. (CVE-2003-0078) 4132111147Snectar 4133111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 4134111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 4135111147Snectar Martin Vuagnoux (EPFL, Ilion)] 4136111147Snectar 4137109998Smarkm Changes between 0.9.6g and 0.9.6h [5 Dec 2002] 4138109998Smarkm 4139109998Smarkm *) New function OPENSSL_cleanse(), which is used to cleanse a section of 4140109998Smarkm memory from it's contents. This is done with a counter that will 4141109998Smarkm place alternating values in each byte. This can be used to solve 4142109998Smarkm two issues: 1) the removal of calls to memset() by highly optimizing 4143109998Smarkm compilers, and 2) cleansing with other values than 0, since those can 4144109998Smarkm be read through on certain media, for example a swap space on disk. 4145109998Smarkm [Geoff Thorpe] 4146109998Smarkm 4147109998Smarkm *) Bugfix: client side session caching did not work with external caching, 4148109998Smarkm because the session->cipher setting was not restored when reloading 4149109998Smarkm from the external cache. This problem was masked, when 4150109998Smarkm SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set. 4151109998Smarkm (Found by Steve Haslam <steve@araqnid.ddts.net>.) 4152109998Smarkm [Lutz Jaenicke] 4153109998Smarkm 4154109998Smarkm *) Fix client_certificate (ssl/s2_clnt.c): The permissible total 4155109998Smarkm length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33. 4156109998Smarkm [Zeev Lieber <zeev-l@yahoo.com>] 4157109998Smarkm 4158109998Smarkm *) Undo an undocumented change introduced in 0.9.6e which caused 4159109998Smarkm repeated calls to OpenSSL_add_all_ciphers() and 4160109998Smarkm OpenSSL_add_all_digests() to be ignored, even after calling 4161109998Smarkm EVP_cleanup(). 4162109998Smarkm [Richard Levitte] 4163109998Smarkm 4164109998Smarkm *) Change the default configuration reader to deal with last line not 4165109998Smarkm being properly terminated. 4166109998Smarkm [Richard Levitte] 4167109998Smarkm 4168109998Smarkm *) Change X509_NAME_cmp() so it applies the special rules on handling 4169109998Smarkm DN values that are of type PrintableString, as well as RDNs of type 4170109998Smarkm emailAddress where the value has the type ia5String. 4171109998Smarkm [stefank@valicert.com via Richard Levitte] 4172109998Smarkm 4173109998Smarkm *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half 4174109998Smarkm the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently 4175109998Smarkm doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be 4176109998Smarkm the bitwise-OR of the two for use by the majority of applications 4177109998Smarkm wanting this behaviour, and update the docs. The documented 4178109998Smarkm behaviour and actual behaviour were inconsistent and had been 4179109998Smarkm changing anyway, so this is more a bug-fix than a behavioural 4180109998Smarkm change. 4181109998Smarkm [Geoff Thorpe, diagnosed by Nadav Har'El] 4182109998Smarkm 4183109998Smarkm *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c 4184109998Smarkm (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes). 4185109998Smarkm [Bodo Moeller] 4186109998Smarkm 4187109998Smarkm *) Fix initialization code race conditions in 4188109998Smarkm SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(), 4189109998Smarkm SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(), 4190109998Smarkm SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(), 4191109998Smarkm TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(), 4192109998Smarkm ssl2_get_cipher_by_char(), 4193109998Smarkm ssl3_get_cipher_by_char(). 4194109998Smarkm [Patrick McCormick <patrick@tellme.com>, Bodo Moeller] 4195109998Smarkm 4196109998Smarkm *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after 4197109998Smarkm the cached sessions are flushed, as the remove_cb() might use ex_data 4198109998Smarkm contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com> 4199109998Smarkm (see [openssl.org #212]). 4200109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 4201109998Smarkm 4202109998Smarkm *) Fix typo in OBJ_txt2obj which incorrectly passed the content 4203109998Smarkm length, instead of the encoding length to d2i_ASN1_OBJECT. 4204109998Smarkm [Steve Henson] 4205109998Smarkm 4206101618Snectar Changes between 0.9.6f and 0.9.6g [9 Aug 2002] 4207101618Snectar 4208101618Snectar *) [In 0.9.6g-engine release:] 4209101618Snectar Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall'). 4210101618Snectar [Lynn Gazis <lgazis@rainbow.com>] 4211101618Snectar 4212101613Snectar Changes between 0.9.6e and 0.9.6f [8 Aug 2002] 4213101613Snectar 4214101613Snectar *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX 4215101613Snectar and get fix the header length calculation. 4216101613Snectar [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>, 4217101613Snectar Alon Kantor <alonk@checkpoint.com> (and others), 4218101613Snectar Steve Henson] 4219101613Snectar 4220101613Snectar *) Use proper error handling instead of 'assertions' in buffer 4221101613Snectar overflow checks added in 0.9.6e. This prevents DoS (the 4222101613Snectar assertions could call abort()). 4223101613Snectar [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller] 4224101613Snectar 4225100936Snectar Changes between 0.9.6d and 0.9.6e [30 Jul 2002] 4226100936Snectar 4227109998Smarkm *) Add various sanity checks to asn1_get_length() to reject 4228109998Smarkm the ASN1 length bytes if they exceed sizeof(long), will appear 4229109998Smarkm negative or the content length exceeds the length of the 4230109998Smarkm supplied buffer. 4231109998Smarkm [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 4232109998Smarkm 4233100936Snectar *) Fix cipher selection routines: ciphers without encryption had no flags 4234100936Snectar for the cipher strength set and where therefore not handled correctly 4235100936Snectar by the selection routines (PR #130). 4236100936Snectar [Lutz Jaenicke] 4237100936Snectar 4238100936Snectar *) Fix EVP_dsa_sha macro. 4239100936Snectar [Nils Larsch] 4240100936Snectar 4241100936Snectar *) New option 4242100936Snectar SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS 4243100936Snectar for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure 4244100936Snectar that was added in OpenSSL 0.9.6d. 4245100936Snectar 4246100936Snectar As the countermeasure turned out to be incompatible with some 4247100936Snectar broken SSL implementations, the new option is part of SSL_OP_ALL. 4248100936Snectar SSL_OP_ALL is usually employed when compatibility with weird SSL 4249100936Snectar implementations is desired (e.g. '-bugs' option to 's_client' and 4250100936Snectar 's_server'), so the new option is automatically set in many 4251100936Snectar applications. 4252100936Snectar [Bodo Moeller] 4253100936Snectar 4254100936Snectar *) Changes in security patch: 4255100936Snectar 4256100936Snectar Changes marked "(CHATS)" were sponsored by the Defense Advanced 4257100936Snectar Research Projects Agency (DARPA) and Air Force Research Laboratory, 4258100936Snectar Air Force Materiel Command, USAF, under agreement number 4259100936Snectar F30602-01-2-0537. 4260100936Snectar 4261100936Snectar *) Add various sanity checks to asn1_get_length() to reject 4262100936Snectar the ASN1 length bytes if they exceed sizeof(long), will appear 4263100936Snectar negative or the content length exceeds the length of the 4264160814Ssimon supplied buffer. (CVE-2002-0659) 4265100936Snectar [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 4266100936Snectar 4267100936Snectar *) Assertions for various potential buffer overflows, not known to 4268100936Snectar happen in practice. 4269100936Snectar [Ben Laurie (CHATS)] 4270100936Snectar 4271100936Snectar *) Various temporary buffers to hold ASCII versions of integers were 4272160814Ssimon too small for 64 bit platforms. (CVE-2002-0655) 4273100936Snectar [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)> 4274100936Snectar 4275100936Snectar *) Remote buffer overflow in SSL3 protocol - an attacker could 4276160814Ssimon supply an oversized session ID to a client. (CVE-2002-0656) 4277100936Snectar [Ben Laurie (CHATS)] 4278100936Snectar 4279100936Snectar *) Remote buffer overflow in SSL2 protocol - an attacker could 4280160814Ssimon supply an oversized client master key. (CVE-2002-0656) 4281100936Snectar [Ben Laurie (CHATS)] 4282100936Snectar 4283100928Snectar Changes between 0.9.6c and 0.9.6d [9 May 2002] 4284100928Snectar 4285100928Snectar *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not 4286100928Snectar encoded as NULL) with id-dsa-with-sha1. 4287100928Snectar [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller] 4288100928Snectar 4289100928Snectar *) Check various X509_...() return values in apps/req.c. 4290100928Snectar [Nils Larsch <nla@trustcenter.de>] 4291100928Snectar 4292100928Snectar *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: 4293100928Snectar an end-of-file condition would erronously be flagged, when the CRLF 4294100928Snectar was just at the end of a processed block. The bug was discovered when 4295100928Snectar processing data through a buffering memory BIO handing the data to a 4296100928Snectar BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov 4297100928Snectar <ptsekov@syntrex.com> and Nedelcho Stanev. 4298100928Snectar [Lutz Jaenicke] 4299100928Snectar 4300100928Snectar *) Implement a countermeasure against a vulnerability recently found 4301100928Snectar in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment 4302100928Snectar before application data chunks to avoid the use of known IVs 4303100928Snectar with data potentially chosen by the attacker. 4304100928Snectar [Bodo Moeller] 4305100928Snectar 4306100928Snectar *) Fix length checks in ssl3_get_client_hello(). 4307100928Snectar [Bodo Moeller] 4308100928Snectar 4309100928Snectar *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently 4310100928Snectar to prevent ssl3_read_internal() from incorrectly assuming that 4311100928Snectar ssl3_read_bytes() found application data while handshake 4312100928Snectar processing was enabled when in fact s->s3->in_read_app_data was 4313100928Snectar merely automatically cleared during the initial handshake. 4314100928Snectar [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>] 4315100928Snectar 4316100928Snectar *) Fix object definitions for Private and Enterprise: they were not 4317100928Snectar recognized in their shortname (=lowercase) representation. Extend 4318100928Snectar obj_dat.pl to issue an error when using undefined keywords instead 4319100928Snectar of silently ignoring the problem (Svenning Sorensen 4320100928Snectar <sss@sss.dnsalias.net>). 4321100928Snectar [Lutz Jaenicke] 4322100928Snectar 4323100928Snectar *) Fix DH_generate_parameters() so that it works for 'non-standard' 4324100928Snectar generators, i.e. generators other than 2 and 5. (Previously, the 4325100928Snectar code did not properly initialise the 'add' and 'rem' values to 4326100928Snectar BN_generate_prime().) 4327100928Snectar 4328100928Snectar In the new general case, we do not insist that 'generator' is 4329100928Snectar actually a primitive root: This requirement is rather pointless; 4330100928Snectar a generator of the order-q subgroup is just as good, if not 4331100928Snectar better. 4332100928Snectar [Bodo Moeller] 4333100928Snectar 4334100928Snectar *) Map new X509 verification errors to alerts. Discovered and submitted by 4335100928Snectar Tom Wu <tom@arcot.com>. 4336100928Snectar [Lutz Jaenicke] 4337100928Snectar 4338100928Snectar *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from 4339100928Snectar returning non-zero before the data has been completely received 4340100928Snectar when using non-blocking I/O. 4341100928Snectar [Bodo Moeller; problem pointed out by John Hughes] 4342100928Snectar 4343100928Snectar *) Some of the ciphers missed the strength entry (SSL_LOW etc). 4344100928Snectar [Ben Laurie, Lutz Jaenicke] 4345100928Snectar 4346100928Snectar *) Fix bug in SSL_clear(): bad sessions were not removed (found by 4347100928Snectar Yoram Zahavi <YoramZ@gilian.com>). 4348100928Snectar [Lutz Jaenicke] 4349100928Snectar 4350100928Snectar *) Add information about CygWin 1.3 and on, and preserve proper 4351100928Snectar configuration for the versions before that. 4352100928Snectar [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 4353100928Snectar 4354100928Snectar *) Make removal from session cache (SSL_CTX_remove_session()) more robust: 4355100928Snectar check whether we deal with a copy of a session and do not delete from 4356100928Snectar the cache in this case. Problem reported by "Izhar Shoshani Levi" 4357100928Snectar <izhar@checkpoint.com>. 4358100928Snectar [Lutz Jaenicke] 4359100928Snectar 4360100928Snectar *) Do not store session data into the internal session cache, if it 4361100928Snectar is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP 4362100928Snectar flag is set). Proposed by Aslam <aslam@funk.com>. 4363100928Snectar [Lutz Jaenicke] 4364100928Snectar 4365100928Snectar *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested 4366100928Snectar value is 0. 4367100928Snectar [Richard Levitte] 4368100928Snectar 4369109998Smarkm *) [In 0.9.6d-engine release:] 4370109998Smarkm Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 4371100928Snectar [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 4372100928Snectar 4373100928Snectar *) Add the configuration target linux-s390x. 4374100928Snectar [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte] 4375100928Snectar 4376100928Snectar *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of 4377100928Snectar ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag 4378100928Snectar variable as an indication that a ClientHello message has been 4379100928Snectar received. As the flag value will be lost between multiple 4380100928Snectar invocations of ssl3_accept when using non-blocking I/O, the 4381100928Snectar function may not be aware that a handshake has actually taken 4382100928Snectar place, thus preventing a new session from being added to the 4383100928Snectar session cache. 4384100928Snectar 4385100928Snectar To avoid this problem, we now set s->new_session to 2 instead of 4386100928Snectar using a local variable. 4387100928Snectar [Lutz Jaenicke, Bodo Moeller] 4388100928Snectar 4389100928Snectar *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) 4390100928Snectar if the SSL_R_LENGTH_MISMATCH error is detected. 4391100928Snectar [Geoff Thorpe, Bodo Moeller] 4392100928Snectar 4393100928Snectar *) New 'shared_ldflag' column in Configure platform table. 4394100928Snectar [Richard Levitte] 4395100928Snectar 4396100928Snectar *) Fix EVP_CIPHER_mode macro. 4397100928Snectar ["Dan S. Camper" <dan@bti.net>] 4398100928Snectar 4399100928Snectar *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown 4400100928Snectar type, we must throw them away by setting rr->length to 0. 4401100928Snectar [D P Chang <dpc@qualys.com>] 4402100928Snectar 440389837Skris Changes between 0.9.6b and 0.9.6c [21 dec 2001] 440489837Skris 440589837Skris *) Fix BN_rand_range bug pointed out by Dominikus Scherkl 440689837Skris <Dominikus.Scherkl@biodata.com>. (The previous implementation 440789837Skris worked incorrectly for those cases where range = 10..._2 and 440889837Skris 3*range is two bits longer than range.) 440989837Skris [Bodo Moeller] 441089837Skris 441189837Skris *) Only add signing time to PKCS7 structures if it is not already 441289837Skris present. 441389837Skris [Steve Henson] 441489837Skris 441589837Skris *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 441689837Skris OBJ_ld_ce should be OBJ_id_ce. 441789837Skris Also some ip-pda OIDs in crypto/objects/objects.txt were 441889837Skris incorrect (cf. RFC 3039). 441989837Skris [Matt Cooper, Frederic Giudicelli, Bodo Moeller] 442089837Skris 442189837Skris *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 442289837Skris returns early because it has nothing to do. 442389837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 442489837Skris 442589837Skris *) [In 0.9.6c-engine release:] 442689837Skris Fix mutex callback return values in crypto/engine/hw_ncipher.c. 442789837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 442889837Skris 442989837Skris *) [In 0.9.6c-engine release:] 443089837Skris Add support for Cryptographic Appliance's keyserver technology. 443189837Skris (Use engine 'keyclient') 443289837Skris [Cryptographic Appliances and Geoff Thorpe] 443389837Skris 443489837Skris *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' 443589837Skris is called via tools/c89.sh because arguments have to be 443689837Skris rearranged (all '-L' options must appear before the first object 443789837Skris modules). 443889837Skris [Richard Shapiro <rshapiro@abinitio.com>] 443989837Skris 444089837Skris *) [In 0.9.6c-engine release:] 444189837Skris Add support for Broadcom crypto accelerator cards, backported 444289837Skris from 0.9.7. 444389837Skris [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] 444489837Skris 444589837Skris *) [In 0.9.6c-engine release:] 444689837Skris Add support for SureWare crypto accelerator cards from 444789837Skris Baltimore Technologies. (Use engine 'sureware') 444889837Skris [Baltimore Technologies and Mark Cox] 444989837Skris 445089837Skris *) [In 0.9.6c-engine release:] 445189837Skris Add support for crypto accelerator cards from Accelerated 445289837Skris Encryption Processing, www.aep.ie. (Use engine 'aep') 445389837Skris [AEP Inc. and Mark Cox] 445489837Skris 445589837Skris *) Add a configuration entry for gcc on UnixWare. 445689837Skris [Gary Benson <gbenson@redhat.com>] 445789837Skris 445889837Skris *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 445989837Skris messages are stored in a single piece (fixed-length part and 446089837Skris variable-length part combined) and fix various bugs found on the way. 446189837Skris [Bodo Moeller] 446289837Skris 446389837Skris *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() 446489837Skris instead. BIO_gethostbyname() does not know what timeouts are 446589837Skris appropriate, so entries would stay in cache even when they have 446689837Skris become invalid. 446789837Skris [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> 446889837Skris 446989837Skris *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 447089837Skris faced with a pathologically small ClientHello fragment that does 447189837Skris not contain client_version: Instead of aborting with an error, 447289837Skris simply choose the highest available protocol version (i.e., 447389837Skris TLS 1.0 unless it is disabled). In practice, ClientHello 447489837Skris messages are never sent like this, but this change gives us 447589837Skris strictly correct behaviour at least for TLS. 447689837Skris [Bodo Moeller] 447789837Skris 447889837Skris *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() 447989837Skris never resets s->method to s->ctx->method when called from within 448089837Skris one of the SSL handshake functions. 448189837Skris [Bodo Moeller; problem pointed out by Niko Baric] 448289837Skris 448389837Skris *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 448489837Skris (sent using the client's version number) if client_version is 448589837Skris smaller than the protocol version in use. Also change 448689837Skris ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 448789837Skris the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 448889837Skris the client will at least see that alert. 448989837Skris [Bodo Moeller] 449089837Skris 449189837Skris *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 449289837Skris correctly. 449389837Skris [Bodo Moeller] 449489837Skris 449589837Skris *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 449689837Skris client receives HelloRequest while in a handshake. 449789837Skris [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] 449889837Skris 449989837Skris *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 450089837Skris should end in 'break', not 'goto end' which circuments various 450189837Skris cleanups done in state SSL_ST_OK. But session related stuff 450289837Skris must be disabled for SSL_ST_OK in the case that we just sent a 450389837Skris HelloRequest. 450489837Skris 450589837Skris Also avoid some overhead by not calling ssl_init_wbio_buffer() 450689837Skris before just sending a HelloRequest. 450789837Skris [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] 450889837Skris 450989837Skris *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 451089837Skris reveal whether illegal block cipher padding was found or a MAC 451189837Skris verification error occured. (Neither SSLerr() codes nor alerts 451289837Skris are directly visible to potential attackers, but the information 451389837Skris may leak via logfiles.) 451489837Skris 451589837Skris Similar changes are not required for the SSL 2.0 implementation 451689837Skris because the number of padding bytes is sent in clear for SSL 2.0, 451789837Skris and the extra bytes are just ignored. However ssl/s2_pkt.c 451889837Skris failed to verify that the purported number of padding bytes is in 451989837Skris the legal range. 452089837Skris [Bodo Moeller] 452189837Skris 452289837Skris *) Add OpenUNIX-8 support including shared libraries 452389837Skris (Boyd Lynn Gerber <gerberb@zenez.com>). 452489837Skris [Lutz Jaenicke] 452589837Skris 452689837Skris *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 452789837Skris 'wristwatch attack' using huge encoding parameters (cf. 452889837Skris James H. Manger's CRYPTO 2001 paper). Note that the 452989837Skris RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 453089837Skris encoding parameters and hence was not vulnerable. 453189837Skris [Bodo Moeller] 453289837Skris 453389837Skris *) BN_sqr() bug fix. 453489837Skris [Ulf M�ller, reported by Jim Ellis <jim.ellis@cavium.com>] 453589837Skris 453689837Skris *) Rabin-Miller test analyses assume uniformly distributed witnesses, 453789837Skris so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 453889837Skris followed by modular reduction. 453989837Skris [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] 454089837Skris 454189837Skris *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 454289837Skris equivalent based on BN_pseudo_rand() instead of BN_rand(). 454389837Skris [Bodo Moeller] 454489837Skris 454589837Skris *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 454689837Skris This function was broken, as the check for a new client hello message 454789837Skris to handle SGC did not allow these large messages. 454889837Skris (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 454989837Skris [Lutz Jaenicke] 455089837Skris 455189837Skris *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). 455289837Skris [Lutz Jaenicke] 455389837Skris 455489837Skris *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 455589837Skris for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 455689837Skris [Lutz Jaenicke] 455789837Skris 455889837Skris *) Rework the configuration and shared library support for Tru64 Unix. 455989837Skris The configuration part makes use of modern compiler features and 456089837Skris still retains old compiler behavior for those that run older versions 456189837Skris of the OS. The shared library support part includes a variant that 456289837Skris uses the RPATH feature, and is available through the special 456389837Skris configuration target "alpha-cc-rpath", which will never be selected 456489837Skris automatically. 456589837Skris [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] 456689837Skris 456789837Skris *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 456889837Skris with the same message size as in ssl3_get_certificate_request(). 456989837Skris Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 457089837Skris messages might inadvertently be reject as too long. 457189837Skris [Petr Lampa <lampa@fee.vutbr.cz>] 457289837Skris 457389837Skris *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 457489837Skris [Andy Polyakov] 457589837Skris 457689837Skris *) Modified SSL library such that the verify_callback that has been set 457789837Skris specificly for an SSL object with SSL_set_verify() is actually being 457889837Skris used. Before the change, a verify_callback set with this function was 457989837Skris ignored and the verify_callback() set in the SSL_CTX at the time of 458089837Skris the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 458189837Skris to allow the necessary settings. 458289837Skris [Lutz Jaenicke] 458389837Skris 458489837Skris *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 458589837Skris explicitly to NULL, as at least on Solaris 8 this seems not always to be 458689837Skris done automatically (in contradiction to the requirements of the C 458789837Skris standard). This made problems when used from OpenSSH. 458889837Skris [Lutz Jaenicke] 458989837Skris 459089837Skris *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 459189837Skris dh->length and always used 459289837Skris 459389837Skris BN_rand_range(priv_key, dh->p). 459489837Skris 459589837Skris BN_rand_range() is not necessary for Diffie-Hellman, and this 459689837Skris specific range makes Diffie-Hellman unnecessarily inefficient if 459789837Skris dh->length (recommended exponent length) is much smaller than the 459889837Skris length of dh->p. We could use BN_rand_range() if the order of 459989837Skris the subgroup was stored in the DH structure, but we only have 460089837Skris dh->length. 460189837Skris 460289837Skris So switch back to 460389837Skris 460489837Skris BN_rand(priv_key, l, ...) 460589837Skris 460689837Skris where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 460789837Skris otherwise. 460889837Skris [Bodo Moeller] 460989837Skris 461089837Skris *) In 461189837Skris 461289837Skris RSA_eay_public_encrypt 461389837Skris RSA_eay_private_decrypt 461489837Skris RSA_eay_private_encrypt (signing) 461589837Skris RSA_eay_public_decrypt (signature verification) 461689837Skris 461789837Skris (default implementations for RSA_public_encrypt, 461889837Skris RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 461989837Skris always reject numbers >= n. 462089837Skris [Bodo Moeller] 462189837Skris 462289837Skris *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 462389837Skris to synchronize access to 'locking_thread'. This is necessary on 462489837Skris systems where access to 'locking_thread' (an 'unsigned long' 462589837Skris variable) is not atomic. 462689837Skris [Bodo Moeller] 462789837Skris 462889837Skris *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 462989837Skris *before* setting the 'crypto_lock_rand' flag. The previous code had 463089837Skris a race condition if 0 is a valid thread ID. 463189837Skris [Travis Vitek <vitek@roguewave.com>] 463289837Skris 463389837Skris *) Add support for shared libraries under Irix. 463489837Skris [Albert Chin-A-Young <china@thewrittenword.com>] 463589837Skris 463689837Skris *) Add configuration option to build on Linux on both big-endian and 463789837Skris little-endian MIPS. 463889837Skris [Ralf Baechle <ralf@uni-koblenz.de>] 463989837Skris 464089837Skris *) Add the possibility to create shared libraries on HP-UX. 464189837Skris [Richard Levitte] 464289837Skris 464379998Skris Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 464479998Skris 464579998Skris *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 464679998Skris to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 464779998Skris Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 464879998Skris PRNG state recovery was possible based on the output of 464979998Skris one PRNG request appropriately sized to gain knowledge on 465079998Skris 'md' followed by enough consecutive 1-byte PRNG requests 465179998Skris to traverse all of 'state'. 465279998Skris 465379998Skris 1. When updating 'md_local' (the current thread's copy of 'md') 465479998Skris during PRNG output generation, hash all of the previous 465579998Skris 'md_local' value, not just the half used for PRNG output. 465679998Skris 465779998Skris 2. Make the number of bytes from 'state' included into the hash 465879998Skris independent from the number of PRNG bytes requested. 465979998Skris 466079998Skris The first measure alone would be sufficient to avoid 466179998Skris Markku-Juhani's attack. (Actually it had never occurred 466279998Skris to me that the half of 'md_local' used for chaining was the 466379998Skris half from which PRNG output bytes were taken -- I had always 466479998Skris assumed that the secret half would be used.) The second 466579998Skris measure makes sure that additional data from 'state' is never 466679998Skris mixed into 'md_local' in small portions; this heuristically 466779998Skris further strengthens the PRNG. 466879998Skris [Bodo Moeller] 466979998Skris 467079998Skris *) Fix crypto/bn/asm/mips3.s. 467179998Skris [Andy Polyakov] 467279998Skris 467379998Skris *) When only the key is given to "enc", the IV is undefined. Print out 467479998Skris an error message in this case. 467579998Skris [Lutz Jaenicke] 467679998Skris 467779998Skris *) Handle special case when X509_NAME is empty in X509 printing routines. 467879998Skris [Steve Henson] 467979998Skris 468079998Skris *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 468179998Skris positive and less than q. 468279998Skris [Bodo Moeller] 468379998Skris 468479998Skris *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 468579998Skris used: it isn't thread safe and the add_lock_callback should handle 468679998Skris that itself. 468779998Skris [Paul Rose <Paul.Rose@bridge.com>] 468879998Skris 468979998Skris *) Verify that incoming data obeys the block size in 469079998Skris ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 469179998Skris [Bodo Moeller] 469279998Skris 469379998Skris *) Fix OAEP check. 469479998Skris [Ulf M�ller, Bodo M�ller] 469579998Skris 469679998Skris *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 469789837Skris RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 469879998Skris when fixing the server behaviour for backwards-compatible 'client 469979998Skris hello' messages. (Note that the attack is impractical against 470079998Skris SSL 3.0 and TLS 1.0 anyway because length and version checking 470179998Skris means that the probability of guessing a valid ciphertext is 470279998Skris around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 470379998Skris paper.) 470479998Skris 470579998Skris Before 0.9.5, the countermeasure (hide the error by generating a 470679998Skris random 'decryption result') did not work properly because 470779998Skris ERR_clear_error() was missing, meaning that SSL_get_error() would 470879998Skris detect the supposedly ignored error. 470979998Skris 471079998Skris Both problems are now fixed. 471179998Skris [Bodo Moeller] 471279998Skris 471379998Skris *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 471479998Skris (previously it was 1024). 471579998Skris [Bodo Moeller] 471679998Skris 471779998Skris *) Fix for compatibility mode trust settings: ignore trust settings 471879998Skris unless some valid trust or reject settings are present. 471979998Skris [Steve Henson] 472079998Skris 472179998Skris *) Fix for blowfish EVP: its a variable length cipher. 472279998Skris [Steve Henson] 472379998Skris 472479998Skris *) Fix various bugs related to DSA S/MIME verification. Handle missing 472579998Skris parameters in DSA public key structures and return an error in the 472679998Skris DSA routines if parameters are absent. 472779998Skris [Steve Henson] 472879998Skris 472979998Skris *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 473079998Skris in the current directory if neither $RANDFILE nor $HOME was set. 473179998Skris RAND_file_name() in 0.9.6a returned NULL in this case. This has 473279998Skris caused some confusion to Windows users who haven't defined $HOME. 473379998Skris Thus RAND_file_name() is changed again: e_os.h can define a 473479998Skris DEFAULT_HOME, which will be used if $HOME is not set. 473579998Skris For Windows, we use "C:"; on other platforms, we still require 473679998Skris environment variables. 473779998Skris 473879998Skris *) Move 'if (!initialized) RAND_poll()' into regions protected by 473979998Skris CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 474079998Skris having multiple threads call RAND_poll() concurrently. 474179998Skris [Bodo Moeller] 474279998Skris 474379998Skris *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 474479998Skris combination of a flag and a thread ID variable. 474579998Skris Otherwise while one thread is in ssleay_rand_bytes (which sets the 474679998Skris flag), *other* threads can enter ssleay_add_bytes without obeying 474789837Skris the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 474879998Skris that they do not hold after the first thread unsets add_do_not_lock). 474979998Skris [Bodo Moeller] 475079998Skris 475179998Skris *) Change bctest again: '-x' expressions are not available in all 475279998Skris versions of 'test'. 475379998Skris [Bodo Moeller] 475479998Skris 475576866Skris Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 475672613Skris 475776866Skris *) Fix a couple of memory leaks in PKCS7_dataDecode() 475876866Skris [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 475976866Skris 476076866Skris *) Change Configure and Makefiles to provide EXE_EXT, which will contain 476176866Skris the default extension for executables, if any. Also, make the perl 476276866Skris scripts that use symlink() to test if it really exists and use "cp" 476376866Skris if it doesn't. All this made OpenSSL compilable and installable in 476476866Skris CygWin. 476576866Skris [Richard Levitte] 476676866Skris 476776866Skris *) Fix for asn1_GetSequence() for indefinite length constructed data. 476876866Skris If SEQUENCE is length is indefinite just set c->slen to the total 476976866Skris amount of data available. 477076866Skris [Steve Henson, reported by shige@FreeBSD.org] 477176866Skris [This change does not apply to 0.9.7.] 477276866Skris 477376866Skris *) Change bctest to avoid here-documents inside command substitution 477476866Skris (workaround for FreeBSD /bin/sh bug). 477576866Skris For compatibility with Ultrix, avoid shell functions (introduced 477676866Skris in the bctest version that searches along $PATH). 477776866Skris [Bodo Moeller] 477876866Skris 477976866Skris *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 478076866Skris with des_encrypt() defined on some operating systems, like Solaris 478176866Skris and UnixWare. 478276866Skris [Richard Levitte] 478376866Skris 478476866Skris *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 478576866Skris On the Importance of Eliminating Errors in Cryptographic 478676866Skris Computations, J. Cryptology 14 (2001) 2, 101-119, 478776866Skris http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 478876866Skris [Ulf Moeller] 478976866Skris 479076866Skris *) MIPS assembler BIGNUM division bug fix. 479176866Skris [Andy Polyakov] 479276866Skris 479376866Skris *) Disabled incorrect Alpha assembler code. 479476866Skris [Richard Levitte] 479576866Skris 479676866Skris *) Fix PKCS#7 decode routines so they correctly update the length 479776866Skris after reading an EOC for the EXPLICIT tag. 479876866Skris [Steve Henson] 479976866Skris [This change does not apply to 0.9.7.] 480076866Skris 480176866Skris *) Fix bug in PKCS#12 key generation routines. This was triggered 480276866Skris if a 3DES key was generated with a 0 initial byte. Include 480376866Skris PKCS12_BROKEN_KEYGEN compilation option to retain the old 480476866Skris (but broken) behaviour. 480576866Skris [Steve Henson] 480676866Skris 480776866Skris *) Enhance bctest to search for a working bc along $PATH and print 480876866Skris it when found. 480976866Skris [Tim Rice <tim@multitalents.net> via Richard Levitte] 481076866Skris 481176866Skris *) Fix memory leaks in err.c: free err_data string if necessary; 481276866Skris don't write to the wrong index in ERR_set_error_data. 481376866Skris [Bodo Moeller] 481476866Skris 481576866Skris *) Implement ssl23_peek (analogous to ssl23_read), which previously 481676866Skris did not exist. 481776866Skris [Bodo Moeller] 481876866Skris 481976866Skris *) Replace rdtsc with _emit statements for VC++ version 5. 482076866Skris [Jeremy Cooper <jeremy@baymoo.org>] 482176866Skris 482276866Skris *) Make it possible to reuse SSLv2 sessions. 482376866Skris [Richard Levitte] 482476866Skris 482576866Skris *) In copy_email() check for >= 0 as a return value for 482676866Skris X509_NAME_get_index_by_NID() since 0 is a valid index. 482776866Skris [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 482876866Skris 482976866Skris *) Avoid coredump with unsupported or invalid public keys by checking if 483076866Skris X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 483176866Skris PKCS7_verify() fails with non detached data. 483276866Skris [Steve Henson] 483376866Skris 483476866Skris *) Don't use getenv in library functions when run as setuid/setgid. 483576866Skris New function OPENSSL_issetugid(). 483676866Skris [Ulf Moeller] 483776866Skris 483876866Skris *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 483976866Skris due to incorrect handling of multi-threading: 484076866Skris 484176866Skris 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 484276866Skris 484376866Skris 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 484476866Skris 484576866Skris 3. Count how many times MemCheck_off() has been called so that 484676866Skris nested use can be treated correctly. This also avoids 484776866Skris inband-signalling in the previous code (which relied on the 484876866Skris assumption that thread ID 0 is impossible). 484976866Skris [Bodo Moeller] 485076866Skris 485176866Skris *) Add "-rand" option also to s_client and s_server. 485276866Skris [Lutz Jaenicke] 485376866Skris 485476866Skris *) Fix CPU detection on Irix 6.x. 485576866Skris [Kurt Hockenbury <khockenb@stevens-tech.edu> and 485676866Skris "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 485776866Skris 485876866Skris *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 485976866Skris was empty. 486076866Skris [Steve Henson] 486176866Skris [This change does not apply to 0.9.7.] 486276866Skris 486376866Skris *) Use the cached encoding of an X509_NAME structure rather than 486476866Skris copying it. This is apparently the reason for the libsafe "errors" 486576866Skris but the code is actually correct. 486676866Skris [Steve Henson] 486776866Skris 486872613Skris *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 486972613Skris Bleichenbacher's DSA attack. 487076866Skris Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 487176866Skris to be set and top=0 forces the highest bit to be set; top=-1 is new 487276866Skris and leaves the highest bit random. 487376866Skris [Ulf Moeller, Bodo Moeller] 487472613Skris 487572613Skris *) In the NCONF_...-based implementations for CONF_... queries 487672613Skris (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 487772613Skris a temporary CONF structure with the data component set to NULL 487872613Skris (which gives segmentation faults in lh_retrieve). 487972613Skris Instead, use NULL for the CONF pointer in CONF_get_string and 488072613Skris CONF_get_number (which may use environment variables) and directly 488172613Skris return NULL from CONF_get_section. 488272613Skris [Bodo Moeller] 488372613Skris 488472613Skris *) Fix potential buffer overrun for EBCDIC. 488572613Skris [Ulf Moeller] 488672613Skris 488772613Skris *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 488872613Skris keyUsage if basicConstraints absent for a CA. 488972613Skris [Steve Henson] 489072613Skris 489172613Skris *) Make SMIME_write_PKCS7() write mail header values with a format that 489272613Skris is more generally accepted (no spaces before the semicolon), since 489372613Skris some programs can't parse those values properly otherwise. Also make 489472613Skris sure BIO's that break lines after each write do not create invalid 489572613Skris headers. 489672613Skris [Richard Levitte] 489772613Skris 489872613Skris *) Make the CRL encoding routines work with empty SEQUENCE OF. The 489972613Skris macros previously used would not encode an empty SEQUENCE OF 490072613Skris and break the signature. 490172613Skris [Steve Henson] 490276866Skris [This change does not apply to 0.9.7.] 490372613Skris 490472613Skris *) Zero the premaster secret after deriving the master secret in 490572613Skris DH ciphersuites. 490672613Skris [Steve Henson] 490772613Skris 490872613Skris *) Add some EVP_add_digest_alias registrations (as found in 490972613Skris OpenSSL_add_all_digests()) to SSL_library_init() 491072613Skris aka OpenSSL_add_ssl_algorithms(). This provides improved 491172613Skris compatibility with peers using X.509 certificates 491272613Skris with unconventional AlgorithmIdentifier OIDs. 491372613Skris [Bodo Moeller] 491472613Skris 491572613Skris *) Fix for Irix with NO_ASM. 491672613Skris ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 491772613Skris 491872613Skris *) ./config script fixes. 491972613Skris [Ulf Moeller, Richard Levitte] 492072613Skris 492172613Skris *) Fix 'openssl passwd -1'. 492272613Skris [Bodo Moeller] 492372613Skris 492472613Skris *) Change PKCS12_key_gen_asc() so it can cope with non null 492572613Skris terminated strings whose length is passed in the passlen 492672613Skris parameter, for example from PEM callbacks. This was done 492772613Skris by adding an extra length parameter to asc2uni(). 492872613Skris [Steve Henson, reported by <oddissey@samsung.co.kr>] 492972613Skris 493072613Skris *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 493172613Skris call failed, free the DSA structure. 493272613Skris [Bodo Moeller] 493372613Skris 493472613Skris *) Fix to uni2asc() to cope with zero length Unicode strings. 493572613Skris These are present in some PKCS#12 files. 493672613Skris [Steve Henson] 493772613Skris 493872613Skris *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 493972613Skris Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 494072613Skris when writing a 32767 byte record. 494172613Skris [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 494272613Skris 494372613Skris *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 494472613Skris obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 494572613Skris 494672613Skris (RSA objects have a reference count access to which is protected 494772613Skris by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 494872613Skris so they are meant to be shared between threads.) 494972613Skris [Bodo Moeller, Geoff Thorpe; original patch submitted by 495072613Skris "Reddie, Steven" <Steven.Reddie@ca.com>] 495172613Skris 495272613Skris *) Fix a deadlock in CRYPTO_mem_leaks(). 495372613Skris [Bodo Moeller] 495472613Skris 495576866Skris *) Use better test patterns in bntest. 495676866Skris [Ulf M�ller] 495776866Skris 495872613Skris *) rand_win.c fix for Borland C. 495972613Skris [Ulf M�ller] 496072613Skris 496172613Skris *) BN_rshift bugfix for n == 0. 496272613Skris [Bodo Moeller] 496372613Skris 496476866Skris *) Add a 'bctest' script that checks for some known 'bc' bugs 496576866Skris so that 'make test' does not abort just because 'bc' is broken. 496676866Skris [Bodo Moeller] 496776866Skris 496872613Skris *) Store verify_result within SSL_SESSION also for client side to 496972613Skris avoid potential security hole. (Re-used sessions on the client side 497072613Skris always resulted in verify_result==X509_V_OK, not using the original 497172613Skris result of the server certificate verification.) 497272613Skris [Lutz Jaenicke] 497372613Skris 497472613Skris *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 497572613Skris SSL3_RT_APPLICATION_DATA, return 0. 497672613Skris Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 497772613Skris [Bodo Moeller] 497872613Skris 497972613Skris *) Fix SSL_peek: 498072613Skris Both ssl2_peek and ssl3_peek, which were totally broken in earlier 498172613Skris releases, have been re-implemented by renaming the previous 498272613Skris implementations of ssl2_read and ssl3_read to ssl2_read_internal 498372613Skris and ssl3_read_internal, respectively, and adding 'peek' parameters 498472613Skris to them. The new ssl[23]_{read,peek} functions are calls to 498572613Skris ssl[23]_read_internal with the 'peek' flag set appropriately. 498672613Skris A 'peek' parameter has also been added to ssl3_read_bytes, which 498772613Skris does the actual work for ssl3_read_internal. 498872613Skris [Bodo Moeller] 498972613Skris 499076866Skris *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 499176866Skris the method-specific "init()" handler. Also clean up ex_data after 499276866Skris calling the method-specific "finish()" handler. Previously, this was 499376866Skris happening the other way round. 499476866Skris [Geoff Thorpe] 499576866Skris 499672613Skris *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 499772613Skris The previous value, 12, was not always sufficient for BN_mod_exp(). 499872613Skris [Bodo Moeller] 499972613Skris 500076866Skris *) Make sure that shared libraries get the internal name engine with 500176866Skris the full version number and not just 0. This should mark the 500276866Skris shared libraries as not backward compatible. Of course, this should 500376866Skris be changed again when we can guarantee backward binary compatibility. 500476866Skris [Richard Levitte] 500576866Skris 500672613Skris *) Fix typo in get_cert_by_subject() in by_dir.c 500772613Skris [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 500872613Skris 500976866Skris *) Rework the system to generate shared libraries: 501076866Skris 501176866Skris - Make note of the expected extension for the shared libraries and 501276866Skris if there is a need for symbolic links from for example libcrypto.so.0 501376866Skris to libcrypto.so.0.9.7. There is extended info in Configure for 501476866Skris that. 501576866Skris 501676866Skris - Make as few rebuilds of the shared libraries as possible. 501776866Skris 501876866Skris - Still avoid linking the OpenSSL programs with the shared libraries. 501976866Skris 502076866Skris - When installing, install the shared libraries separately from the 502176866Skris static ones. 502276866Skris [Richard Levitte] 502376866Skris 502472613Skris *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 502572613Skris 502672613Skris Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 502772613Skris and not in SSL_clear because the latter is also used by the 502872613Skris accept/connect functions; previously, the settings made by 502972613Skris SSL_set_read_ahead would be lost during the handshake. 503072613Skris [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 503172613Skris 503272613Skris *) Correct util/mkdef.pl to be selective about disabled algorithms. 503372613Skris Previously, it would create entries for disableed algorithms no 503472613Skris matter what. 503572613Skris [Richard Levitte] 503672613Skris 503772613Skris *) Added several new manual pages for SSL_* function. 503872613Skris [Lutz Jaenicke] 503972613Skris 504068651Skris Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 504168651Skris 504268651Skris *) In ssl23_get_client_hello, generate an error message when faced 504368651Skris with an initial SSL 3.0/TLS record that is too small to contain the 504468651Skris first two bytes of the ClientHello message, i.e. client_version. 504568651Skris (Note that this is a pathologic case that probably has never happened 504668651Skris in real life.) The previous approach was to use the version number 504768651Skris from the record header as a substitute; but our protocol choice 504868651Skris should not depend on that one because it is not authenticated 504968651Skris by the Finished messages. 505068651Skris [Bodo Moeller] 505168651Skris 505268651Skris *) More robust randomness gathering functions for Windows. 505368651Skris [Jeffrey Altman <jaltman@columbia.edu>] 505468651Skris 505568651Skris *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 505668651Skris not set then we don't setup the error code for issuer check errors 505768651Skris to avoid possibly overwriting other errors which the callback does 505868651Skris handle. If an application does set the flag then we assume it knows 505968651Skris what it is doing and can handle the new informational codes 506068651Skris appropriately. 506168651Skris [Steve Henson] 506268651Skris 506368651Skris *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 506468651Skris a general "ANY" type, as such it should be able to decode anything 506568651Skris including tagged types. However it didn't check the class so it would 506668651Skris wrongly interpret tagged types in the same way as their universal 506768651Skris counterpart and unknown types were just rejected. Changed so that the 506868651Skris tagged and unknown types are handled in the same way as a SEQUENCE: 506968651Skris that is the encoding is stored intact. There is also a new type 507068651Skris "V_ASN1_OTHER" which is used when the class is not universal, in this 507168651Skris case we have no idea what the actual type is so we just lump them all 507268651Skris together. 507368651Skris [Steve Henson] 507468651Skris 507568651Skris *) On VMS, stdout may very well lead to a file that is written to 507668651Skris in a record-oriented fashion. That means that every write() will 507768651Skris write a separate record, which will be read separately by the 507868651Skris programs trying to read from it. This can be very confusing. 507968651Skris 508068651Skris The solution is to put a BIO filter in the way that will buffer 508168651Skris text until a linefeed is reached, and then write everything a 508268651Skris line at a time, so every record written will be an actual line, 508368651Skris not chunks of lines and not (usually doesn't happen, but I've 508468651Skris seen it once) several lines in one record. BIO_f_linebuffer() is 508568651Skris the answer. 508668651Skris 508768651Skris Currently, it's a VMS-only method, because that's where it has 508868651Skris been tested well enough. 508968651Skris [Richard Levitte] 509068651Skris 509168651Skris *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 509268651Skris it can return incorrect results. 509368651Skris (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 509468651Skris but it was in 0.9.6-beta[12].) 509568651Skris [Bodo Moeller] 509668651Skris 509768651Skris *) Disable the check for content being present when verifying detached 509868651Skris signatures in pk7_smime.c. Some versions of Netscape (wrongly) 509968651Skris include zero length content when signing messages. 510068651Skris [Steve Henson] 510168651Skris 510268651Skris *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 510368651Skris BIO_ctrl (for BIO pairs). 510468651Skris [Bodo M�ller] 510568651Skris 510668651Skris *) Add DSO method for VMS. 510768651Skris [Richard Levitte] 510868651Skris 510968651Skris *) Bug fix: Montgomery multiplication could produce results with the 511068651Skris wrong sign. 511168651Skris [Ulf M�ller] 511268651Skris 511368651Skris *) Add RPM specification openssl.spec and modify it to build three 511468651Skris packages. The default package contains applications, application 511568651Skris documentation and run-time libraries. The devel package contains 511668651Skris include files, static libraries and function documentation. The 511768651Skris doc package contains the contents of the doc directory. The original 511868651Skris openssl.spec was provided by Damien Miller <djm@mindrot.org>. 511968651Skris [Richard Levitte] 512068651Skris 512168651Skris *) Add a large number of documentation files for many SSL routines. 512268651Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 512368651Skris 512468651Skris *) Add a configuration entry for Sony News 4. 512568651Skris [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 512668651Skris 512768651Skris *) Don't set the two most significant bits to one when generating a 512868651Skris random number < q in the DSA library. 512968651Skris [Ulf M�ller] 513068651Skris 513168651Skris *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 513268651Skris behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 513368651Skris the underlying transport is blocking) if a handshake took place. 513468651Skris (The default behaviour is needed by applications such as s_client 513568651Skris and s_server that use select() to determine when to use SSL_read; 513668651Skris but for applications that know in advance when to expect data, it 513768651Skris just makes things more complicated.) 513868651Skris [Bodo Moeller] 513968651Skris 514068651Skris *) Add RAND_egd_bytes(), which gives control over the number of bytes read 514168651Skris from EGD. 514268651Skris [Ben Laurie] 514368651Skris 514468651Skris *) Add a few more EBCDIC conditionals that make `req' and `x509' 514568651Skris work better on such systems. 514668651Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 514768651Skris 514868651Skris *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 514968651Skris Update PKCS12_parse() so it copies the friendlyName and the 515068651Skris keyid to the certificates aux info. 515168651Skris [Steve Henson] 515268651Skris 515368651Skris *) Fix bug in PKCS7_verify() which caused an infinite loop 515468651Skris if there was more than one signature. 515568651Skris [Sven Uszpelkat <su@celocom.de>] 515668651Skris 515768651Skris *) Major change in util/mkdef.pl to include extra information 515868651Skris about each symbol, as well as presentig variables as well 515968651Skris as functions. This change means that there's n more need 516068651Skris to rebuild the .num files when some algorithms are excluded. 516168651Skris [Richard Levitte] 516268651Skris 516368651Skris *) Allow the verify time to be set by an application, 516468651Skris rather than always using the current time. 516568651Skris [Steve Henson] 516668651Skris 516768651Skris *) Phase 2 verify code reorganisation. The certificate 516868651Skris verify code now looks up an issuer certificate by a 516968651Skris number of criteria: subject name, authority key id 517068651Skris and key usage. It also verifies self signed certificates 517168651Skris by the same criteria. The main comparison function is 517268651Skris X509_check_issued() which performs these checks. 517368651Skris 517468651Skris Lot of changes were necessary in order to support this 517568651Skris without completely rewriting the lookup code. 517668651Skris 517768651Skris Authority and subject key identifier are now cached. 517868651Skris 517968651Skris The LHASH 'certs' is X509_STORE has now been replaced 518068651Skris by a STACK_OF(X509_OBJECT). This is mainly because an 518168651Skris LHASH can't store or retrieve multiple objects with 518268651Skris the same hash value. 518368651Skris 518468651Skris As a result various functions (which were all internal 518568651Skris use only) have changed to handle the new X509_STORE 518668651Skris structure. This will break anything that messed round 518768651Skris with X509_STORE internally. 518868651Skris 518968651Skris The functions X509_STORE_add_cert() now checks for an 519068651Skris exact match, rather than just subject name. 519168651Skris 519268651Skris The X509_STORE API doesn't directly support the retrieval 519368651Skris of multiple certificates matching a given criteria, however 519468651Skris this can be worked round by performing a lookup first 519568651Skris (which will fill the cache with candidate certificates) 519668651Skris and then examining the cache for matches. This is probably 519768651Skris the best we can do without throwing out X509_LOOKUP 519868651Skris entirely (maybe later...). 519968651Skris 520068651Skris The X509_VERIFY_CTX structure has been enhanced considerably. 520168651Skris 520268651Skris All certificate lookup operations now go via a get_issuer() 520368651Skris callback. Although this currently uses an X509_STORE it 520468651Skris can be replaced by custom lookups. This is a simple way 520568651Skris to bypass the X509_STORE hackery necessary to make this 520668651Skris work and makes it possible to use more efficient techniques 520768651Skris in future. A very simple version which uses a simple 520868651Skris STACK for its trusted certificate store is also provided 520968651Skris using X509_STORE_CTX_trusted_stack(). 521068651Skris 521168651Skris The verify_cb() and verify() callbacks now have equivalents 521268651Skris in the X509_STORE_CTX structure. 521368651Skris 521468651Skris X509_STORE_CTX also has a 'flags' field which can be used 521568651Skris to customise the verify behaviour. 521668651Skris [Steve Henson] 521768651Skris 521868651Skris *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 521968651Skris excludes S/MIME capabilities. 522068651Skris [Steve Henson] 522168651Skris 522268651Skris *) When a certificate request is read in keep a copy of the 522368651Skris original encoding of the signed data and use it when outputing 522468651Skris again. Signatures then use the original encoding rather than 522568651Skris a decoded, encoded version which may cause problems if the 522668651Skris request is improperly encoded. 522768651Skris [Steve Henson] 522868651Skris 522968651Skris *) For consistency with other BIO_puts implementations, call 523068651Skris buffer_write(b, ...) directly in buffer_puts instead of calling 523168651Skris BIO_write(b, ...). 523268651Skris 523368651Skris In BIO_puts, increment b->num_write as in BIO_write. 523468651Skris [Peter.Sylvester@EdelWeb.fr] 523568651Skris 523668651Skris *) Fix BN_mul_word for the case where the word is 0. (We have to use 523768651Skris BN_zero, we may not return a BIGNUM with an array consisting of 523868651Skris words set to zero.) 523968651Skris [Bodo Moeller] 524068651Skris 524168651Skris *) Avoid calling abort() from within the library when problems are 524268651Skris detected, except if preprocessor symbols have been defined 524368651Skris (such as REF_CHECK, BN_DEBUG etc.). 524468651Skris [Bodo Moeller] 524568651Skris 524668651Skris *) New openssl application 'rsautl'. This utility can be 524768651Skris used for low level RSA operations. DER public key 524868651Skris BIO/fp routines also added. 524968651Skris [Steve Henson] 525068651Skris 525168651Skris *) New Configure entry and patches for compiling on QNX 4. 525268651Skris [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 525368651Skris 525468651Skris *) A demo state-machine implementation was sponsored by 525568651Skris Nuron (http://www.nuron.com/) and is now available in 525668651Skris demos/state_machine. 525768651Skris [Ben Laurie] 525868651Skris 525968651Skris *) New options added to the 'dgst' utility for signature 526068651Skris generation and verification. 526168651Skris [Steve Henson] 526268651Skris 526368651Skris *) Unrecognized PKCS#7 content types are now handled via a 526468651Skris catch all ASN1_TYPE structure. This allows unsupported 526568651Skris types to be stored as a "blob" and an application can 526668651Skris encode and decode it manually. 526768651Skris [Steve Henson] 526868651Skris 526968651Skris *) Fix various signed/unsigned issues to make a_strex.c 527068651Skris compile under VC++. 527168651Skris [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 527268651Skris 527368651Skris *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 527468651Skris length if passed a buffer. ASN1_INTEGER_to_BN failed 527568651Skris if passed a NULL BN and its argument was negative. 527668651Skris [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 527768651Skris 527868651Skris *) Modification to PKCS#7 encoding routines to output definite 527968651Skris length encoding. Since currently the whole structures are in 528068651Skris memory there's not real point in using indefinite length 528168651Skris constructed encoding. However if OpenSSL is compiled with 528268651Skris the flag PKCS7_INDEFINITE_ENCODING the old form is used. 528368651Skris [Steve Henson] 528468651Skris 528568651Skris *) Added BIO_vprintf() and BIO_vsnprintf(). 528668651Skris [Richard Levitte] 528768651Skris 528868651Skris *) Added more prefixes to parse for in the the strings written 528968651Skris through a logging bio, to cover all the levels that are available 529068651Skris through syslog. The prefixes are now: 529168651Skris 529268651Skris PANIC, EMERG, EMR => LOG_EMERG 529368651Skris ALERT, ALR => LOG_ALERT 529468651Skris CRIT, CRI => LOG_CRIT 529568651Skris ERROR, ERR => LOG_ERR 529668651Skris WARNING, WARN, WAR => LOG_WARNING 529768651Skris NOTICE, NOTE, NOT => LOG_NOTICE 529868651Skris INFO, INF => LOG_INFO 529968651Skris DEBUG, DBG => LOG_DEBUG 530068651Skris 530168651Skris and as before, if none of those prefixes are present at the 530268651Skris beginning of the string, LOG_ERR is chosen. 530368651Skris 530468651Skris On Win32, the LOG_* levels are mapped according to this: 530568651Skris 530668651Skris LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 530768651Skris LOG_WARNING => EVENTLOG_WARNING_TYPE 530868651Skris LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 530968651Skris 531068651Skris [Richard Levitte] 531168651Skris 531268651Skris *) Made it possible to reconfigure with just the configuration 531368651Skris argument "reconf" or "reconfigure". The command line arguments 531468651Skris are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 531568651Skris and are retrieved from there when reconfiguring. 531668651Skris [Richard Levitte] 531768651Skris 531868651Skris *) MD4 implemented. 531968651Skris [Assar Westerlund <assar@sics.se>, Richard Levitte] 532068651Skris 532168651Skris *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 532268651Skris [Richard Levitte] 532368651Skris 532468651Skris *) The obj_dat.pl script was messing up the sorting of object 532568651Skris names. The reason was that it compared the quoted version 532668651Skris of strings as a result "OCSP" > "OCSP Signing" because 532768651Skris " > SPACE. Changed script to store unquoted versions of 532868651Skris names and add quotes on output. It was also omitting some 532968651Skris names from the lookup table if they were given a default 533068651Skris value (that is if SN is missing it is given the same 533168651Skris value as LN and vice versa), these are now added on the 533268651Skris grounds that if an object has a name we should be able to 533368651Skris look it up. Finally added warning output when duplicate 533468651Skris short or long names are found. 533568651Skris [Steve Henson] 533668651Skris 533768651Skris *) Changes needed for Tandem NSK. 533868651Skris [Scott Uroff <scott@xypro.com>] 533968651Skris 534068651Skris *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 534168651Skris RSA_padding_check_SSLv23(), special padding was never detected 534268651Skris and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 534368651Skris version rollback attacks was not effective. 534468651Skris 534568651Skris In s23_clnt.c, don't use special rollback-attack detection padding 534668651Skris (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 534768651Skris client; similarly, in s23_srvr.c, don't do the rollback check if 534868651Skris SSL 2.0 is the only protocol enabled in the server. 534968651Skris [Bodo Moeller] 535068651Skris 535168651Skris *) Make it possible to get hexdumps of unprintable data with 'openssl 535268651Skris asn1parse'. By implication, the functions ASN1_parse_dump() and 535368651Skris BIO_dump_indent() are added. 535468651Skris [Richard Levitte] 535568651Skris 535668651Skris *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 535768651Skris these print out strings and name structures based on various 535868651Skris flags including RFC2253 support and proper handling of 535968651Skris multibyte characters. Added options to the 'x509' utility 536068651Skris to allow the various flags to be set. 536168651Skris [Steve Henson] 536268651Skris 536368651Skris *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 536468651Skris Also change the functions X509_cmp_current_time() and 536568651Skris X509_gmtime_adj() work with an ASN1_TIME structure, 536668651Skris this will enable certificates using GeneralizedTime in validity 536768651Skris dates to be checked. 536868651Skris [Steve Henson] 536968651Skris 537068651Skris *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 537168651Skris negative public key encodings) on by default, 537268651Skris NO_NEG_PUBKEY_BUG can be set to disable it. 537368651Skris [Steve Henson] 537468651Skris 537568651Skris *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 537668651Skris content octets. An i2c_ASN1_OBJECT is unnecessary because 537768651Skris the encoding can be trivially obtained from the structure. 537868651Skris [Steve Henson] 537968651Skris 538068651Skris *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 538168651Skris not read locks (CRYPTO_r_[un]lock). 538268651Skris [Bodo Moeller] 538368651Skris 538468651Skris *) A first attempt at creating official support for shared 538568651Skris libraries through configuration. I've kept it so the 538668651Skris default is static libraries only, and the OpenSSL programs 538768651Skris are always statically linked for now, but there are 538868651Skris preparations for dynamic linking in place. 538989837Skris This has been tested on Linux and Tru64. 539068651Skris [Richard Levitte] 539168651Skris 539268651Skris *) Randomness polling function for Win9x, as described in: 539368651Skris Peter Gutmann, Software Generation of Practically Strong 539468651Skris Random Numbers. 539568651Skris [Ulf M�ller] 539668651Skris 539768651Skris *) Fix so PRNG is seeded in req if using an already existing 539868651Skris DSA key. 539968651Skris [Steve Henson] 540068651Skris 540168651Skris *) New options to smime application. -inform and -outform 540268651Skris allow alternative formats for the S/MIME message including 540368651Skris PEM and DER. The -content option allows the content to be 540468651Skris specified separately. This should allow things like Netscape 540568651Skris form signing output easier to verify. 540668651Skris [Steve Henson] 540768651Skris 540868651Skris *) Fix the ASN1 encoding of tags using the 'long form'. 540968651Skris [Steve Henson] 541068651Skris 541168651Skris *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 541268651Skris STRING types. These convert content octets to and from the 541368651Skris underlying type. The actual tag and length octets are 541468651Skris already assumed to have been read in and checked. These 541568651Skris are needed because all other string types have virtually 541668651Skris identical handling apart from the tag. By having versions 541768651Skris of the ASN1 functions that just operate on content octets 541868651Skris IMPLICIT tagging can be handled properly. It also allows 541968651Skris the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 542068651Skris and ASN1_INTEGER are identical apart from the tag. 542168651Skris [Steve Henson] 542268651Skris 542368651Skris *) Change the handling of OID objects as follows: 542468651Skris 542568651Skris - New object identifiers are inserted in objects.txt, following 542668651Skris the syntax given in objects.README. 542768651Skris - objects.pl is used to process obj_mac.num and create a new 542868651Skris obj_mac.h. 542968651Skris - obj_dat.pl is used to create a new obj_dat.h, using the data in 543068651Skris obj_mac.h. 543168651Skris 543268651Skris This is currently kind of a hack, and the perl code in objects.pl 543368651Skris isn't very elegant, but it works as I intended. The simplest way 543468651Skris to check that it worked correctly is to look in obj_dat.h and 543568651Skris check the array nid_objs and make sure the objects haven't moved 543668651Skris around (this is important!). Additions are OK, as well as 543768651Skris consistent name changes. 543868651Skris [Richard Levitte] 543968651Skris 544068651Skris *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 544168651Skris [Bodo Moeller] 544268651Skris 544368651Skris *) Addition of the command line parameter '-rand file' to 'openssl req'. 544468651Skris The given file adds to whatever has already been seeded into the 544568651Skris random pool through the RANDFILE configuration file option or 544668651Skris environment variable, or the default random state file. 544768651Skris [Richard Levitte] 544868651Skris 544968651Skris *) mkstack.pl now sorts each macro group into lexical order. 545068651Skris Previously the output order depended on the order the files 545168651Skris appeared in the directory, resulting in needless rewriting 545268651Skris of safestack.h . 545368651Skris [Steve Henson] 545468651Skris 545568651Skris *) Patches to make OpenSSL compile under Win32 again. Mostly 545668651Skris work arounds for the VC++ problem that it treats func() as 545768651Skris func(void). Also stripped out the parts of mkdef.pl that 545868651Skris added extra typesafe functions: these no longer exist. 545968651Skris [Steve Henson] 546068651Skris 546168651Skris *) Reorganisation of the stack code. The macros are now all 546268651Skris collected in safestack.h . Each macro is defined in terms of 546368651Skris a "stack macro" of the form SKM_<name>(type, a, b). The 546468651Skris DEBUG_SAFESTACK is now handled in terms of function casts, 546568651Skris this has the advantage of retaining type safety without the 546668651Skris use of additional functions. If DEBUG_SAFESTACK is not defined 546768651Skris then the non typesafe macros are used instead. Also modified the 546868651Skris mkstack.pl script to handle the new form. Needs testing to see 546968651Skris if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 547068651Skris the default if no major problems. Similar behaviour for ASN1_SET_OF 547168651Skris and PKCS12_STACK_OF. 547268651Skris [Steve Henson] 547368651Skris 547468651Skris *) When some versions of IIS use the 'NET' form of private key the 547568651Skris key derivation algorithm is different. Normally MD5(password) is 547668651Skris used as a 128 bit RC4 key. In the modified case 547768651Skris MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 547868651Skris new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 547968651Skris as the old Netscape_RSA functions except they have an additional 548068651Skris 'sgckey' parameter which uses the modified algorithm. Also added 548168651Skris an -sgckey command line option to the rsa utility. Thanks to 548268651Skris Adrian Peck <bertie@ncipher.com> for posting details of the modified 548368651Skris algorithm to openssl-dev. 548468651Skris [Steve Henson] 548568651Skris 548668651Skris *) The evp_local.h macros were using 'c.##kname' which resulted in 548768651Skris invalid expansion on some systems (SCO 5.0.5 for example). 548868651Skris Corrected to 'c.kname'. 548968651Skris [Phillip Porch <root@theporch.com>] 549068651Skris 549168651Skris *) New X509_get1_email() and X509_REQ_get1_email() functions that return 549268651Skris a STACK of email addresses from a certificate or request, these look 549368651Skris in the subject name and the subject alternative name extensions and 549468651Skris omit any duplicate addresses. 549568651Skris [Steve Henson] 549668651Skris 549768651Skris *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 549868651Skris This makes DSA verification about 2 % faster. 549968651Skris [Bodo Moeller] 550068651Skris 550168651Skris *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 550268651Skris (meaning that now 2^5 values will be precomputed, which is only 4 KB 550368651Skris plus overhead for 1024 bit moduli). 550468651Skris This makes exponentiations about 0.5 % faster for 1024 bit 550568651Skris exponents (as measured by "openssl speed rsa2048"). 550668651Skris [Bodo Moeller] 550768651Skris 550868651Skris *) Rename memory handling macros to avoid conflicts with other 550968651Skris software: 551068651Skris Malloc => OPENSSL_malloc 551168651Skris Malloc_locked => OPENSSL_malloc_locked 551268651Skris Realloc => OPENSSL_realloc 551368651Skris Free => OPENSSL_free 551468651Skris [Richard Levitte] 551568651Skris 551668651Skris *) New function BN_mod_exp_mont_word for small bases (roughly 15% 551768651Skris faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 551868651Skris [Bodo Moeller] 551968651Skris 552068651Skris *) CygWin32 support. 552168651Skris [John Jarvie <jjarvie@newsguy.com>] 552268651Skris 552368651Skris *) The type-safe stack code has been rejigged. It is now only compiled 552468651Skris in when OpenSSL is configured with the DEBUG_SAFESTACK option and 552568651Skris by default all type-specific stack functions are "#define"d back to 552668651Skris standard stack functions. This results in more streamlined output 552768651Skris but retains the type-safety checking possibilities of the original 552868651Skris approach. 552968651Skris [Geoff Thorpe] 553068651Skris 553168651Skris *) The STACK code has been cleaned up, and certain type declarations 553268651Skris that didn't make a lot of sense have been brought in line. This has 553368651Skris also involved a cleanup of sorts in safestack.h to more correctly 553468651Skris map type-safe stack functions onto their plain stack counterparts. 553568651Skris This work has also resulted in a variety of "const"ifications of 553668651Skris lots of the code, especially "_cmp" operations which should normally 553768651Skris be prototyped with "const" parameters anyway. 553868651Skris [Geoff Thorpe] 553968651Skris 554068651Skris *) When generating bytes for the first time in md_rand.c, 'stir the pool' 554168651Skris by seeding with STATE_SIZE dummy bytes (with zero entropy count). 554268651Skris (The PRNG state consists of two parts, the large pool 'state' and 'md', 554368651Skris where all of 'md' is used each time the PRNG is used, but 'state' 554468651Skris is used only indexed by a cyclic counter. As entropy may not be 554568651Skris well distributed from the beginning, 'md' is important as a 554668651Skris chaining variable. However, the output function chains only half 554768651Skris of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 554868651Skris all of 'md', and seeding with STATE_SIZE dummy bytes will result 554968651Skris in all of 'state' being rewritten, with the new values depending 555068651Skris on virtually all of 'md'. This overcomes the 80 bit limitation.) 555168651Skris [Bodo Moeller] 555268651Skris 555368651Skris *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 555468651Skris the handshake is continued after ssl_verify_cert_chain(); 555568651Skris otherwise, if SSL_VERIFY_NONE is set, remaining error codes 555668651Skris can lead to 'unexplainable' connection aborts later. 555768651Skris [Bodo Moeller; problem tracked down by Lutz Jaenicke] 555868651Skris 555968651Skris *) Major EVP API cipher revision. 556068651Skris Add hooks for extra EVP features. This allows various cipher 556168651Skris parameters to be set in the EVP interface. Support added for variable 556268651Skris key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 556368651Skris setting of RC2 and RC5 parameters. 556468651Skris 556568651Skris Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 556668651Skris ciphers. 556768651Skris 556868651Skris Remove lots of duplicated code from the EVP library. For example *every* 556968651Skris cipher init() function handles the 'iv' in the same way according to the 557068651Skris cipher mode. They also all do nothing if the 'key' parameter is NULL and 557168651Skris for CFB and OFB modes they zero ctx->num. 557268651Skris 557368651Skris New functionality allows removal of S/MIME code RC2 hack. 557468651Skris 557568651Skris Most of the routines have the same form and so can be declared in terms 557668651Skris of macros. 557768651Skris 557868651Skris By shifting this to the top level EVP_CipherInit() it can be removed from 557968651Skris all individual ciphers. If the cipher wants to handle IVs or keys 558068651Skris differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 558168651Skris flags. 558268651Skris 558368651Skris Change lots of functions like EVP_EncryptUpdate() to now return a 558468651Skris value: although software versions of the algorithms cannot fail 558568651Skris any installed hardware versions can. 558668651Skris [Steve Henson] 558768651Skris 558868651Skris *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 558968651Skris this option is set, tolerate broken clients that send the negotiated 559068651Skris protocol version number instead of the requested protocol version 559168651Skris number. 559268651Skris [Bodo Moeller] 559368651Skris 559468651Skris *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 559568651Skris i.e. non-zero for export ciphersuites, zero otherwise. 559668651Skris Previous versions had this flag inverted, inconsistent with 559768651Skris rsa_tmp_cb (..._TMP_RSA_CB). 559868651Skris [Bodo Moeller; problem reported by Amit Chopra] 559968651Skris 560068651Skris *) Add missing DSA library text string. Work around for some IIS 560168651Skris key files with invalid SEQUENCE encoding. 560268651Skris [Steve Henson] 560368651Skris 560468651Skris *) Add a document (doc/standards.txt) that list all kinds of standards 560568651Skris and so on that are implemented in OpenSSL. 560668651Skris [Richard Levitte] 560768651Skris 560868651Skris *) Enhance c_rehash script. Old version would mishandle certificates 560968651Skris with the same subject name hash and wouldn't handle CRLs at all. 561068651Skris Added -fingerprint option to crl utility, to support new c_rehash 561168651Skris features. 561268651Skris [Steve Henson] 561368651Skris 561468651Skris *) Eliminate non-ANSI declarations in crypto.h and stack.h. 561568651Skris [Ulf M�ller] 561668651Skris 561768651Skris *) Fix for SSL server purpose checking. Server checking was 561868651Skris rejecting certificates which had extended key usage present 561968651Skris but no ssl client purpose. 562068651Skris [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 562168651Skris 562268651Skris *) Make PKCS#12 code work with no password. The PKCS#12 spec 562368651Skris is a little unclear about how a blank password is handled. 562468651Skris Since the password in encoded as a BMPString with terminating 562568651Skris double NULL a zero length password would end up as just the 562668651Skris double NULL. However no password at all is different and is 562768651Skris handled differently in the PKCS#12 key generation code. NS 562868651Skris treats a blank password as zero length. MSIE treats it as no 562968651Skris password on export: but it will try both on import. We now do 563068651Skris the same: PKCS12_parse() tries zero length and no password if 563168651Skris the password is set to "" or NULL (NULL is now a valid password: 563268651Skris it wasn't before) as does the pkcs12 application. 563368651Skris [Steve Henson] 563468651Skris 563568651Skris *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 563668651Skris perror when PEM_read_bio_X509_REQ fails, the error message must 563768651Skris be obtained from the error queue. 563868651Skris [Bodo Moeller] 563968651Skris 564068651Skris *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 564168651Skris it in ERR_remove_state if appropriate, and change ERR_get_state 564268651Skris accordingly to avoid race conditions (this is necessary because 564368651Skris thread_hash is no longer constant once set). 564468651Skris [Bodo Moeller] 564568651Skris 564668651Skris *) Bugfix for linux-elf makefile.one. 564768651Skris [Ulf M�ller] 564868651Skris 564968651Skris *) RSA_get_default_method() will now cause a default 565068651Skris RSA_METHOD to be chosen if one doesn't exist already. 565168651Skris Previously this was only set during a call to RSA_new() 565268651Skris or RSA_new_method(NULL) meaning it was possible for 565368651Skris RSA_get_default_method() to return NULL. 565468651Skris [Geoff Thorpe] 565568651Skris 565668651Skris *) Added native name translation to the existing DSO code 565768651Skris that will convert (if the flag to do so is set) filenames 565868651Skris that are sufficiently small and have no path information 565968651Skris into a canonical native form. Eg. "blah" converted to 566068651Skris "libblah.so" or "blah.dll" etc. 566168651Skris [Geoff Thorpe] 566268651Skris 566368651Skris *) New function ERR_error_string_n(e, buf, len) which is like 566468651Skris ERR_error_string(e, buf), but writes at most 'len' bytes 566568651Skris including the 0 terminator. For ERR_error_string_n, 'buf' 566668651Skris may not be NULL. 566768651Skris [Damien Miller <djm@mindrot.org>, Bodo Moeller] 566868651Skris 566968651Skris *) CONF library reworked to become more general. A new CONF 567068651Skris configuration file reader "class" is implemented as well as a 567168651Skris new functions (NCONF_*, for "New CONF") to handle it. The now 567268651Skris old CONF_* functions are still there, but are reimplemented to 567368651Skris work in terms of the new functions. Also, a set of functions 567468651Skris to handle the internal storage of the configuration data is 567568651Skris provided to make it easier to write new configuration file 567668651Skris reader "classes" (I can definitely see something reading a 567768651Skris configuration file in XML format, for example), called _CONF_*, 567868651Skris or "the configuration storage API"... 567968651Skris 568068651Skris The new configuration file reading functions are: 568168651Skris 568268651Skris NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 568368651Skris NCONF_get_section, NCONF_get_string, NCONF_get_numbre 568468651Skris 568568651Skris NCONF_default, NCONF_WIN32 568668651Skris 568768651Skris NCONF_dump_fp, NCONF_dump_bio 568868651Skris 568968651Skris NCONF_default and NCONF_WIN32 are method (or "class") choosers, 569068651Skris NCONF_new creates a new CONF object. This works in the same way 569168651Skris as other interfaces in OpenSSL, like the BIO interface. 569268651Skris NCONF_dump_* dump the internal storage of the configuration file, 569368651Skris which is useful for debugging. All other functions take the same 569468651Skris arguments as the old CONF_* functions wth the exception of the 569568651Skris first that must be a `CONF *' instead of a `LHASH *'. 569668651Skris 569768651Skris To make it easer to use the new classes with the old CONF_* functions, 569868651Skris the function CONF_set_default_method is provided. 569968651Skris [Richard Levitte] 570068651Skris 570168651Skris *) Add '-tls1' option to 'openssl ciphers', which was already 570268651Skris mentioned in the documentation but had not been implemented. 570368651Skris (This option is not yet really useful because even the additional 570468651Skris experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 570568651Skris [Bodo Moeller] 570668651Skris 570768651Skris *) Initial DSO code added into libcrypto for letting OpenSSL (and 570868651Skris OpenSSL-based applications) load shared libraries and bind to 570968651Skris them in a portable way. 571068651Skris [Geoff Thorpe, with contributions from Richard Levitte] 571168651Skris 571259191Skris Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 571359191Skris 571459191Skris *) Make sure _lrotl and _lrotr are only used with MSVC. 571559191Skris 571659191Skris *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 571759191Skris (the default implementation of RAND_status). 571859191Skris 571959191Skris *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 572059191Skris to '-clrext' (= clear extensions), as intended and documented. 572159191Skris [Bodo Moeller; inconsistency pointed out by Michael Attili 572259191Skris <attili@amaxo.com>] 572359191Skris 572459191Skris *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 572559191Skris was larger than the MD block size. 572659191Skris [Steve Henson, pointed out by Yost William <YostW@tce.com>] 572759191Skris 572859191Skris *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 572959191Skris fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 573059191Skris using the passed key: if the passed key was a private key the result 573159191Skris of X509_print(), for example, would be to print out all the private key 573259191Skris components. 573359191Skris [Steve Henson] 573459191Skris 573559191Skris *) des_quad_cksum() byte order bug fix. 573659191Skris [Ulf M�ller, using the problem description in krb4-0.9.7, where 573759191Skris the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 573859191Skris 573959191Skris *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 574059191Skris discouraged. 574159191Skris [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 574259191Skris 574359191Skris *) For easily testing in shell scripts whether some command 574459191Skris 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 574559191Skris returns with exit code 0 iff no command of the given name is available. 574659191Skris 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 574759191Skris the output goes to stdout and nothing is printed to stderr. 574859191Skris Additional arguments are always ignored. 574959191Skris 575059191Skris Since for each cipher there is a command of the same name, 575159191Skris the 'no-cipher' compilation switches can be tested this way. 575259191Skris 575359191Skris ('openssl no-XXX' is not able to detect pseudo-commands such 575459191Skris as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 575559191Skris [Bodo Moeller] 575659191Skris 575759191Skris *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 575859191Skris [Bodo Moeller] 575959191Skris 576059191Skris *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 576159191Skris is set; it will be thrown away anyway because each handshake creates 576259191Skris its own key. 576359191Skris ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 576459191Skris to parameters -- in previous versions (since OpenSSL 0.9.3) the 576559191Skris 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 576659191Skris you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 576759191Skris [Bodo Moeller] 576859191Skris 576959191Skris *) New s_client option -ign_eof: EOF at stdin is ignored, and 577059191Skris 'Q' and 'R' lose their special meanings (quit/renegotiate). 577159191Skris This is part of what -quiet does; unlike -quiet, -ign_eof 577259191Skris does not suppress any output. 577359191Skris [Richard Levitte] 577459191Skris 577559191Skris *) Add compatibility options to the purpose and trust code. The 577659191Skris purpose X509_PURPOSE_ANY is "any purpose" which automatically 577759191Skris accepts a certificate or CA, this was the previous behaviour, 577859191Skris with all the associated security issues. 577959191Skris 578059191Skris X509_TRUST_COMPAT is the old trust behaviour: only and 578159191Skris automatically trust self signed roots in certificate store. A 578259191Skris new trust setting X509_TRUST_DEFAULT is used to specify that 578359191Skris a purpose has no associated trust setting and it should instead 578459191Skris use the value in the default purpose. 578559191Skris [Steve Henson] 578659191Skris 578759191Skris *) Fix the PKCS#8 DSA private key code so it decodes keys again 578859191Skris and fix a memory leak. 578959191Skris [Steve Henson] 579059191Skris 579159191Skris *) In util/mkerr.pl (which implements 'make errors'), preserve 579259191Skris reason strings from the previous version of the .c file, as 579359191Skris the default to have only downcase letters (and digits) in 579459191Skris automatically generated reasons codes is not always appropriate. 579559191Skris [Bodo Moeller] 579659191Skris 579759191Skris *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 579859191Skris using strerror. Previously, ERR_reason_error_string() returned 579959191Skris library names as reason strings for SYSerr; but SYSerr is a special 580059191Skris case where small numbers are errno values, not library numbers. 580159191Skris [Bodo Moeller] 580259191Skris 580359191Skris *) Add '-dsaparam' option to 'openssl dhparam' application. This 580459191Skris converts DSA parameters into DH parameters. (When creating parameters, 580559191Skris DSA_generate_parameters is used.) 580659191Skris [Bodo Moeller] 580759191Skris 580859191Skris *) Include 'length' (recommended exponent length) in C code generated 580959191Skris by 'openssl dhparam -C'. 581059191Skris [Bodo Moeller] 581159191Skris 581259191Skris *) The second argument to set_label in perlasm was already being used 581359191Skris so couldn't be used as a "file scope" flag. Moved to third argument 581459191Skris which was free. 581559191Skris [Steve Henson] 581659191Skris 581759191Skris *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 581859191Skris instead of RAND_bytes for encryption IVs and salts. 581959191Skris [Bodo Moeller] 582059191Skris 582159191Skris *) Include RAND_status() into RAND_METHOD instead of implementing 582259191Skris it only for md_rand.c Otherwise replacing the PRNG by calling 582359191Skris RAND_set_rand_method would be impossible. 582459191Skris [Bodo Moeller] 582559191Skris 582659191Skris *) Don't let DSA_generate_key() enter an infinite loop if the random 582759191Skris number generation fails. 582859191Skris [Bodo Moeller] 582959191Skris 583059191Skris *) New 'rand' application for creating pseudo-random output. 583159191Skris [Bodo Moeller] 583259191Skris 583359191Skris *) Added configuration support for Linux/IA64 583459191Skris [Rolf Haberrecker <rolf@suse.de>] 583559191Skris 583659191Skris *) Assembler module support for Mingw32. 583759191Skris [Ulf M�ller] 583859191Skris 583959191Skris *) Shared library support for HPUX (in shlib/). 584059191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 584159191Skris 584259191Skris *) Shared library support for Solaris gcc. 584359191Skris [Lutz Behnke <behnke@trustcenter.de>] 584459191Skris 584559191Skris Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 584659191Skris 584759191Skris *) PKCS7_encrypt() was adding text MIME headers twice because they 584859191Skris were added manually and by SMIME_crlf_copy(). 584959191Skris [Steve Henson] 585059191Skris 585159191Skris *) In bntest.c don't call BN_rand with zero bits argument. 585259191Skris [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 585359191Skris 585459191Skris *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 585559191Skris case was implemented. This caused BN_div_recp() to fail occasionally. 585659191Skris [Ulf M�ller] 585759191Skris 585859191Skris *) Add an optional second argument to the set_label() in the perl 585959191Skris assembly language builder. If this argument exists and is set 586059191Skris to 1 it signals that the assembler should use a symbol whose 586159191Skris scope is the entire file, not just the current function. This 586259191Skris is needed with MASM which uses the format label:: for this scope. 586359191Skris [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 586459191Skris 586559191Skris *) Change the ASN1 types so they are typedefs by default. Before 586659191Skris almost all types were #define'd to ASN1_STRING which was causing 586759191Skris STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 586859191Skris for example. 586959191Skris [Steve Henson] 587059191Skris 587159191Skris *) Change names of new functions to the new get1/get0 naming 587259191Skris convention: After 'get1', the caller owns a reference count 587359191Skris and has to call ..._free; 'get0' returns a pointer to some 587459191Skris data structure without incrementing reference counters. 587559191Skris (Some of the existing 'get' functions increment a reference 587659191Skris counter, some don't.) 587759191Skris Similarly, 'set1' and 'add1' functions increase reference 587859191Skris counters or duplicate objects. 587959191Skris [Steve Henson] 588059191Skris 588159191Skris *) Allow for the possibility of temp RSA key generation failure: 588259191Skris the code used to assume it always worked and crashed on failure. 588359191Skris [Steve Henson] 588459191Skris 588559191Skris *) Fix potential buffer overrun problem in BIO_printf(). 588659191Skris [Ulf M�ller, using public domain code by Patrick Powell; problem 588759191Skris pointed out by David Sacerdote <das33@cornell.edu>] 588859191Skris 588959191Skris *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 589059191Skris RAND_egd() and RAND_status(). In the command line application, 589159191Skris the EGD socket can be specified like a seed file using RANDFILE 589259191Skris or -rand. 589359191Skris [Ulf M�ller] 589459191Skris 589559191Skris *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 589659191Skris Some CAs (e.g. Verisign) distribute certificates in this form. 589759191Skris [Steve Henson] 589859191Skris 589959191Skris *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 590059191Skris list to exclude them. This means that no special compilation option 590159191Skris is needed to use anonymous DH: it just needs to be included in the 590259191Skris cipher list. 590359191Skris [Steve Henson] 590459191Skris 590559191Skris *) Change the EVP_MD_CTX_type macro so its meaning consistent with 590659191Skris EVP_MD_type. The old functionality is available in a new macro called 590759191Skris EVP_MD_md(). Change code that uses it and update docs. 590859191Skris [Steve Henson] 590959191Skris 591059191Skris *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 591159191Skris where the 'void *' argument is replaced by a function pointer argument. 591259191Skris Previously 'void *' was abused to point to functions, which works on 591359191Skris many platforms, but is not correct. As these functions are usually 591459191Skris called by macros defined in OpenSSL header files, most source code 591559191Skris should work without changes. 591659191Skris [Richard Levitte] 591759191Skris 591859191Skris *) <openssl/opensslconf.h> (which is created by Configure) now contains 591959191Skris sections with information on -D... compiler switches used for 592059191Skris compiling the library so that applications can see them. To enable 592159191Skris one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 592259191Skris must be defined. E.g., 592359191Skris #define OPENSSL_ALGORITHM_DEFINES 592459191Skris #include <openssl/opensslconf.h> 592559191Skris defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 592659191Skris [Richard Levitte, Ulf and Bodo M�ller] 592759191Skris 592859191Skris *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 592959191Skris record layer. 593059191Skris [Bodo Moeller] 593159191Skris 593259191Skris *) Change the 'other' type in certificate aux info to a STACK_OF 593359191Skris X509_ALGOR. Although not an AlgorithmIdentifier as such it has 593459191Skris the required ASN1 format: arbitrary types determined by an OID. 593559191Skris [Steve Henson] 593659191Skris 593759191Skris *) Add some PEM_write_X509_REQ_NEW() functions and a command line 593859191Skris argument to 'req'. This is not because the function is newer or 593959191Skris better than others it just uses the work 'NEW' in the certificate 594059191Skris request header lines. Some software needs this. 594159191Skris [Steve Henson] 594259191Skris 594359191Skris *) Reorganise password command line arguments: now passwords can be 594459191Skris obtained from various sources. Delete the PEM_cb function and make 594559191Skris it the default behaviour: i.e. if the callback is NULL and the 594659191Skris usrdata argument is not NULL interpret it as a null terminated pass 594759191Skris phrase. If usrdata and the callback are NULL then the pass phrase 594859191Skris is prompted for as usual. 594959191Skris [Steve Henson] 595059191Skris 595159191Skris *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 595259191Skris the support is automatically enabled. The resulting binaries will 595359191Skris autodetect the card and use it if present. 595459191Skris [Ben Laurie and Compaq Inc.] 595559191Skris 595659191Skris *) Work around for Netscape hang bug. This sends certificate request 595759191Skris and server done in one record. Since this is perfectly legal in the 595859191Skris SSL/TLS protocol it isn't a "bug" option and is on by default. See 595959191Skris the bugs/SSLv3 entry for more info. 596059191Skris [Steve Henson] 596159191Skris 596259191Skris *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 596359191Skris [Andy Polyakov] 596459191Skris 596559191Skris *) Add -rand argument to smime and pkcs12 applications and read/write 596659191Skris of seed file. 596759191Skris [Steve Henson] 596859191Skris 596959191Skris *) New 'passwd' tool for crypt(3) and apr1 password hashes. 597059191Skris [Bodo Moeller] 597159191Skris 597259191Skris *) Add command line password options to the remaining applications. 597359191Skris [Steve Henson] 597459191Skris 597559191Skris *) Bug fix for BN_div_recp() for numerators with an even number of 597659191Skris bits. 597759191Skris [Ulf M�ller] 597859191Skris 597959191Skris *) More tests in bntest.c, and changed test_bn output. 598059191Skris [Ulf M�ller] 598159191Skris 598259191Skris *) ./config recognizes MacOS X now. 598359191Skris [Andy Polyakov] 598459191Skris 598559191Skris *) Bug fix for BN_div() when the first words of num and divsor are 598659191Skris equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 598759191Skris [Ulf M�ller] 598859191Skris 598959191Skris *) Add support for various broken PKCS#8 formats, and command line 599059191Skris options to produce them. 599159191Skris [Steve Henson] 599259191Skris 599359191Skris *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 599459191Skris get temporary BIGNUMs from a BN_CTX. 599559191Skris [Ulf M�ller] 599659191Skris 599759191Skris *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 599859191Skris for p == 0. 599959191Skris [Ulf M�ller] 600059191Skris 600159191Skris *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 600259191Skris include a #define from the old name to the new. The original intent 600359191Skris was that statically linked binaries could for example just call 600459191Skris SSLeay_add_all_ciphers() to just add ciphers to the table and not 600559191Skris link with digests. This never worked becayse SSLeay_add_all_digests() 600659191Skris and SSLeay_add_all_ciphers() were in the same source file so calling 600759191Skris one would link with the other. They are now in separate source files. 600859191Skris [Steve Henson] 600959191Skris 601059191Skris *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 601159191Skris [Steve Henson] 601259191Skris 601359191Skris *) Use a less unusual form of the Miller-Rabin primality test (it used 601459191Skris a binary algorithm for exponentiation integrated into the Miller-Rabin 601559191Skris loop, our standard modexp algorithms are faster). 601659191Skris [Bodo Moeller] 601759191Skris 601859191Skris *) Support for the EBCDIC character set completed. 601959191Skris [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 602059191Skris 602159191Skris *) Source code cleanups: use const where appropriate, eliminate casts, 602259191Skris use void * instead of char * in lhash. 602359191Skris [Ulf M�ller] 602459191Skris 602559191Skris *) Bugfix: ssl3_send_server_key_exchange was not restartable 602659191Skris (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 602759191Skris this the server could overwrite ephemeral keys that the client 602859191Skris has already seen). 602959191Skris [Bodo Moeller] 603059191Skris 603159191Skris *) Turn DSA_is_prime into a macro that calls BN_is_prime, 603259191Skris using 50 iterations of the Rabin-Miller test. 603359191Skris 603459191Skris DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 603559191Skris iterations of the Rabin-Miller test as required by the appendix 603659191Skris to FIPS PUB 186[-1]) instead of DSA_is_prime. 603759191Skris As BN_is_prime_fasttest includes trial division, DSA parameter 603859191Skris generation becomes much faster. 603959191Skris 604059191Skris This implies a change for the callback functions in DSA_is_prime 604159191Skris and DSA_generate_parameters: The callback function is called once 604259191Skris for each positive witness in the Rabin-Miller test, not just 604359191Skris occasionally in the inner loop; and the parameters to the 604459191Skris callback function now provide an iteration count for the outer 604559191Skris loop rather than for the current invocation of the inner loop. 604659191Skris DSA_generate_parameters additionally can call the callback 604759191Skris function with an 'iteration count' of -1, meaning that a 604859191Skris candidate has passed the trial division test (when q is generated 604959191Skris from an application-provided seed, trial division is skipped). 605059191Skris [Bodo Moeller] 605159191Skris 605259191Skris *) New function BN_is_prime_fasttest that optionally does trial 605359191Skris division before starting the Rabin-Miller test and has 605459191Skris an additional BN_CTX * argument (whereas BN_is_prime always 605559191Skris has to allocate at least one BN_CTX). 605659191Skris 'callback(1, -1, cb_arg)' is called when a number has passed the 605759191Skris trial division stage. 605859191Skris [Bodo Moeller] 605959191Skris 606059191Skris *) Fix for bug in CRL encoding. The validity dates weren't being handled 606159191Skris as ASN1_TIME. 606259191Skris [Steve Henson] 606359191Skris 606459191Skris *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 606559191Skris [Steve Henson] 606659191Skris 606759191Skris *) New function BN_pseudo_rand(). 606859191Skris [Ulf M�ller] 606959191Skris 607059191Skris *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 607159191Skris bignum version of BN_from_montgomery() with the working code from 607259191Skris SSLeay 0.9.0 (the word based version is faster anyway), and clean up 607359191Skris the comments. 607459191Skris [Ulf M�ller] 607559191Skris 607659191Skris *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 607759191Skris made it impossible to use the same SSL_SESSION data structure in 607859191Skris SSL2 clients in multiple threads. 607959191Skris [Bodo Moeller] 608059191Skris 608159191Skris *) The return value of RAND_load_file() no longer counts bytes obtained 608259191Skris by stat(). RAND_load_file(..., -1) is new and uses the complete file 608359191Skris to seed the PRNG (previously an explicit byte count was required). 608459191Skris [Ulf M�ller, Bodo M�ller] 608559191Skris 608659191Skris *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 608759191Skris used (char *) instead of (void *) and had casts all over the place. 608859191Skris [Steve Henson] 608959191Skris 609059191Skris *) Make BN_generate_prime() return NULL on error if ret!=NULL. 609159191Skris [Ulf M�ller] 609259191Skris 609359191Skris *) Retain source code compatibility for BN_prime_checks macro: 609459191Skris BN_is_prime(..., BN_prime_checks, ...) now uses 609559191Skris BN_prime_checks_for_size to determine the appropriate number of 609659191Skris Rabin-Miller iterations. 609759191Skris [Ulf M�ller] 609859191Skris 609959191Skris *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 610059191Skris DH_CHECK_P_NOT_SAFE_PRIME. 610159191Skris (Check if this is true? OpenPGP calls them "strong".) 610259191Skris [Ulf M�ller] 610359191Skris 610459191Skris *) Merge the functionality of "dh" and "gendh" programs into a new program 610559191Skris "dhparam". The old programs are retained for now but will handle DH keys 610659191Skris (instead of parameters) in future. 610759191Skris [Steve Henson] 610859191Skris 610959191Skris *) Make the ciphers, s_server and s_client programs check the return values 611059191Skris when a new cipher list is set. 611159191Skris [Steve Henson] 611259191Skris 611359191Skris *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 611459191Skris ciphers. Before when the 56bit ciphers were enabled the sorting was 611559191Skris wrong. 611659191Skris 611759191Skris The syntax for the cipher sorting has been extended to support sorting by 611859191Skris cipher-strength (using the strength_bits hard coded in the tables). 611959191Skris The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 612059191Skris 612159191Skris Fix a bug in the cipher-command parser: when supplying a cipher command 612259191Skris string with an "undefined" symbol (neither command nor alphanumeric 612359191Skris [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 612459191Skris an error is flagged. 612559191Skris 612659191Skris Due to the strength-sorting extension, the code of the 612759191Skris ssl_create_cipher_list() function was completely rearranged. I hope that 612859191Skris the readability was also increased :-) 612959191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 613059191Skris 613159191Skris *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 613259191Skris for the first serial number and places 2 in the serial number file. This 613359191Skris avoids problems when the root CA is created with serial number zero and 613459191Skris the first user certificate has the same issuer name and serial number 613559191Skris as the root CA. 613659191Skris [Steve Henson] 613759191Skris 613859191Skris *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 613959191Skris the new code. Add documentation for this stuff. 614059191Skris [Steve Henson] 614159191Skris 614259191Skris *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 614359191Skris X509_*() to X509at_*() on the grounds that they don't handle X509 614459191Skris structures and behave in an analagous way to the X509v3 functions: 614559191Skris they shouldn't be called directly but wrapper functions should be used 614659191Skris instead. 614759191Skris 614859191Skris So we also now have some wrapper functions that call the X509at functions 614959191Skris when passed certificate requests. (TO DO: similar things can be done with 615059191Skris PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 615159191Skris things. Some of these need some d2i or i2d and print functionality 615259191Skris because they handle more complex structures.) 615359191Skris [Steve Henson] 615459191Skris 615559191Skris *) Add missing #ifndefs that caused missing symbols when building libssl 615659191Skris as a shared library without RSA. Use #ifndef NO_SSL2 instead of 615759191Skris NO_RSA in ssl/s2*.c. 615859191Skris [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M�ller] 615959191Skris 616059191Skris *) Precautions against using the PRNG uninitialized: RAND_bytes() now 616159191Skris has a return value which indicates the quality of the random data 616259191Skris (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 616359191Skris error queue. New function RAND_pseudo_bytes() generates output that is 616459191Skris guaranteed to be unique but not unpredictable. RAND_add is like 616559191Skris RAND_seed, but takes an extra argument for an entropy estimate 616659191Skris (RAND_seed always assumes full entropy). 616759191Skris [Ulf M�ller] 616859191Skris 616959191Skris *) Do more iterations of Rabin-Miller probable prime test (specifically, 617059191Skris 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 617159191Skris instead of only 2 for all lengths; see BN_prime_checks_for_size definition 617259191Skris in crypto/bn/bn_prime.c for the complete table). This guarantees a 617359191Skris false-positive rate of at most 2^-80 for random input. 617459191Skris [Bodo Moeller] 617559191Skris 617659191Skris *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 617759191Skris [Bodo Moeller] 617859191Skris 617959191Skris *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 618059191Skris in the 0.9.5 release), this returns the chain 618159191Skris from an X509_CTX structure with a dup of the stack and all 618259191Skris the X509 reference counts upped: so the stack will exist 618359191Skris after X509_CTX_cleanup() has been called. Modify pkcs12.c 618459191Skris to use this. 618559191Skris 618659191Skris Also make SSL_SESSION_print() print out the verify return 618759191Skris code. 618859191Skris [Steve Henson] 618959191Skris 619059191Skris *) Add manpage for the pkcs12 command. Also change the default 619159191Skris behaviour so MAC iteration counts are used unless the new 619259191Skris -nomaciter option is used. This improves file security and 619359191Skris only older versions of MSIE (4.0 for example) need it. 619459191Skris [Steve Henson] 619559191Skris 619659191Skris *) Honor the no-xxx Configure options when creating .DEF files. 619759191Skris [Ulf M�ller] 619859191Skris 619959191Skris *) Add PKCS#10 attributes to field table: challengePassword, 620059191Skris unstructuredName and unstructuredAddress. These are taken from 620159191Skris draft PKCS#9 v2.0 but are compatible with v1.2 provided no 620259191Skris international characters are used. 620359191Skris 620459191Skris More changes to X509_ATTRIBUTE code: allow the setting of types 620559191Skris based on strings. Remove the 'loc' parameter when adding 620659191Skris attributes because these will be a SET OF encoding which is sorted 620759191Skris in ASN1 order. 620859191Skris [Steve Henson] 620959191Skris 621059191Skris *) Initial changes to the 'req' utility to allow request generation 621159191Skris automation. This will allow an application to just generate a template 621259191Skris file containing all the field values and have req construct the 621359191Skris request. 621459191Skris 621559191Skris Initial support for X509_ATTRIBUTE handling. Stacks of these are 621659191Skris used all over the place including certificate requests and PKCS#7 621759191Skris structures. They are currently handled manually where necessary with 621859191Skris some primitive wrappers for PKCS#7. The new functions behave in a 621959191Skris manner analogous to the X509 extension functions: they allow 622059191Skris attributes to be looked up by NID and added. 622159191Skris 622259191Skris Later something similar to the X509V3 code would be desirable to 622359191Skris automatically handle the encoding, decoding and printing of the 622459191Skris more complex types. The string types like challengePassword can 622559191Skris be handled by the string table functions. 622659191Skris 622759191Skris Also modified the multi byte string table handling. Now there is 622859191Skris a 'global mask' which masks out certain types. The table itself 622959191Skris can use the flag STABLE_NO_MASK to ignore the mask setting: this 623059191Skris is useful when for example there is only one permissible type 623159191Skris (as in countryName) and using the mask might result in no valid 623259191Skris types at all. 623359191Skris [Steve Henson] 623459191Skris 623559191Skris *) Clean up 'Finished' handling, and add functions SSL_get_finished and 623659191Skris SSL_get_peer_finished to allow applications to obtain the latest 623759191Skris Finished messages sent to the peer or expected from the peer, 623859191Skris respectively. (SSL_get_peer_finished is usually the Finished message 623959191Skris actually received from the peer, otherwise the protocol will be aborted.) 624059191Skris 624159191Skris As the Finished message are message digests of the complete handshake 624259191Skris (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 624359191Skris be used for external authentication procedures when the authentication 624459191Skris provided by SSL/TLS is not desired or is not enough. 624559191Skris [Bodo Moeller] 624659191Skris 624759191Skris *) Enhanced support for Alpha Linux is added. Now ./config checks if 624859191Skris the host supports BWX extension and if Compaq C is present on the 624959191Skris $PATH. Just exploiting of the BWX extension results in 20-30% 625059191Skris performance kick for some algorithms, e.g. DES and RC4 to mention 625159191Skris a couple. Compaq C in turn generates ~20% faster code for MD5 and 625259191Skris SHA1. 625359191Skris [Andy Polyakov] 625459191Skris 625559191Skris *) Add support for MS "fast SGC". This is arguably a violation of the 625659191Skris SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 625759191Skris weak crypto and after checking the certificate is SGC a second one 625859191Skris with strong crypto. MS SGC stops the first handshake after receiving 625959191Skris the server certificate message and sends a second client hello. Since 626059191Skris a server will typically do all the time consuming operations before 626159191Skris expecting any further messages from the client (server key exchange 626259191Skris is the most expensive) there is little difference between the two. 626359191Skris 626459191Skris To get OpenSSL to support MS SGC we have to permit a second client 626559191Skris hello message after we have sent server done. In addition we have to 626659191Skris reset the MAC if we do get this second client hello. 626759191Skris [Steve Henson] 626859191Skris 626959191Skris *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 627059191Skris if a DER encoded private key is RSA or DSA traditional format. Changed 627159191Skris d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 627259191Skris format DER encoded private key. Newer code should use PKCS#8 format which 627359191Skris has the key type encoded in the ASN1 structure. Added DER private key 627459191Skris support to pkcs8 application. 627559191Skris [Steve Henson] 627659191Skris 627759191Skris *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 627859191Skris ciphersuites has been selected (as required by the SSL 3/TLS 1 627959191Skris specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 628059191Skris is set, we interpret this as a request to violate the specification 628159191Skris (the worst that can happen is a handshake failure, and 'correct' 628259191Skris behaviour would result in a handshake failure anyway). 628359191Skris [Bodo Moeller] 628459191Skris 628559191Skris *) In SSL_CTX_add_session, take into account that there might be multiple 628659191Skris SSL_SESSION structures with the same session ID (e.g. when two threads 628759191Skris concurrently obtain them from an external cache). 628859191Skris The internal cache can handle only one SSL_SESSION with a given ID, 628959191Skris so if there's a conflict, we now throw out the old one to achieve 629059191Skris consistency. 629159191Skris [Bodo Moeller] 629259191Skris 629359191Skris *) Add OIDs for idea and blowfish in CBC mode. This will allow both 629459191Skris to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 629559191Skris some routines that use cipher OIDs: some ciphers do not have OIDs 629659191Skris defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 629759191Skris example. 629859191Skris [Steve Henson] 629959191Skris 630059191Skris *) Simplify the trust setting structure and code. Now we just have 630159191Skris two sequences of OIDs for trusted and rejected settings. These will 630259191Skris typically have values the same as the extended key usage extension 630359191Skris and any application specific purposes. 630459191Skris 630559191Skris The trust checking code now has a default behaviour: it will just 630659191Skris check for an object with the same NID as the passed id. Functions can 630759191Skris be provided to override either the default behaviour or the behaviour 630859191Skris for a given id. SSL client, server and email already have functions 630959191Skris in place for compatibility: they check the NID and also return "trusted" 631059191Skris if the certificate is self signed. 631159191Skris [Steve Henson] 631259191Skris 631359191Skris *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 631459191Skris traditional format into an EVP_PKEY structure. 631559191Skris [Steve Henson] 631659191Skris 631759191Skris *) Add a password callback function PEM_cb() which either prompts for 631859191Skris a password if usr_data is NULL or otherwise assumes it is a null 631959191Skris terminated password. Allow passwords to be passed on command line 632059191Skris environment or config files in a few more utilities. 632159191Skris [Steve Henson] 632259191Skris 632359191Skris *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 632459191Skris keys. Add some short names for PKCS#8 PBE algorithms and allow them 632559191Skris to be specified on the command line for the pkcs8 and pkcs12 utilities. 632659191Skris Update documentation. 632759191Skris [Steve Henson] 632859191Skris 632959191Skris *) Support for ASN1 "NULL" type. This could be handled before by using 633059191Skris ASN1_TYPE but there wasn't any function that would try to read a NULL 633159191Skris and produce an error if it couldn't. For compatibility we also have 633259191Skris ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 633359191Skris don't allocate anything because they don't need to. 633459191Skris [Steve Henson] 633559191Skris 633659191Skris *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 633759191Skris for details. 633859191Skris [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 633959191Skris 634059191Skris *) Rebuild of the memory allocation routines used by OpenSSL code and 634159191Skris possibly others as well. The purpose is to make an interface that 634259191Skris provide hooks so anyone can build a separate set of allocation and 634359191Skris deallocation routines to be used by OpenSSL, for example memory 634459191Skris pool implementations, or something else, which was previously hard 634559191Skris since Malloc(), Realloc() and Free() were defined as macros having 634659191Skris the values malloc, realloc and free, respectively (except for Win32 634759191Skris compilations). The same is provided for memory debugging code. 634859191Skris OpenSSL already comes with functionality to find memory leaks, but 634959191Skris this gives people a chance to debug other memory problems. 635059191Skris 635159191Skris With these changes, a new set of functions and macros have appeared: 635259191Skris 635368651Skris CRYPTO_set_mem_debug_functions() [F] 635468651Skris CRYPTO_get_mem_debug_functions() [F] 635568651Skris CRYPTO_dbg_set_options() [F] 635668651Skris CRYPTO_dbg_get_options() [F] 635768651Skris CRYPTO_malloc_debug_init() [M] 635859191Skris 635959191Skris The memory debug functions are NULL by default, unless the library 636059191Skris is compiled with CRYPTO_MDEBUG or friends is defined. If someone 636159191Skris wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 636259191Skris gives the standard debugging functions that come with OpenSSL) or 636359191Skris CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 636459191Skris provided by the library user) must be used. When the standard 636559191Skris debugging functions are used, CRYPTO_dbg_set_options can be used to 636659191Skris request additional information: 636759191Skris CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 636859191Skris the CRYPTO_MDEBUG_xxx macro when compiling the library. 636959191Skris 637059191Skris Also, things like CRYPTO_set_mem_functions will always give the 637159191Skris expected result (the new set of functions is used for allocation 637259191Skris and deallocation) at all times, regardless of platform and compiler 637359191Skris options. 637459191Skris 637559191Skris To finish it up, some functions that were never use in any other 637659191Skris way than through macros have a new API and new semantic: 637759191Skris 637859191Skris CRYPTO_dbg_malloc() 637959191Skris CRYPTO_dbg_realloc() 638059191Skris CRYPTO_dbg_free() 638159191Skris 638259191Skris All macros of value have retained their old syntax. 638359191Skris [Richard Levitte and Bodo Moeller] 638459191Skris 638559191Skris *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 638659191Skris ordering of SMIMECapabilities wasn't in "strength order" and there 638759191Skris was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 638859191Skris algorithm. 638959191Skris [Steve Henson] 639059191Skris 639159191Skris *) Some ASN1 types with illegal zero length encoding (INTEGER, 639259191Skris ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 639359191Skris [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 639459191Skris 639559191Skris *) Merge in my S/MIME library for OpenSSL. This provides a simple 639659191Skris S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 639759191Skris functionality to handle multipart/signed properly) and a utility 639859191Skris called 'smime' to call all this stuff. This is based on code I 639959191Skris originally wrote for Celo who have kindly allowed it to be 640059191Skris included in OpenSSL. 640159191Skris [Steve Henson] 640259191Skris 640359191Skris *) Add variants des_set_key_checked and des_set_key_unchecked of 640459191Skris des_set_key (aka des_key_sched). Global variable des_check_key 640559191Skris decides which of these is called by des_set_key; this way 640659191Skris des_check_key behaves as it always did, but applications and 640759191Skris the library itself, which was buggy for des_check_key == 1, 640859191Skris have a cleaner way to pick the version they need. 640959191Skris [Bodo Moeller] 641059191Skris 641159191Skris *) New function PKCS12_newpass() which changes the password of a 641259191Skris PKCS12 structure. 641359191Skris [Steve Henson] 641459191Skris 641559191Skris *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 641659191Skris dynamic mix. In both cases the ids can be used as an index into the 641759191Skris table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 641859191Skris functions so they accept a list of the field values and the 641959191Skris application doesn't need to directly manipulate the X509_TRUST 642059191Skris structure. 642159191Skris [Steve Henson] 642259191Skris 642359191Skris *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 642459191Skris need initialising. 642559191Skris [Steve Henson] 642659191Skris 642759191Skris *) Modify the way the V3 extension code looks up extensions. This now 642859191Skris works in a similar way to the object code: we have some "standard" 642959191Skris extensions in a static table which is searched with OBJ_bsearch() 643059191Skris and the application can add dynamic ones if needed. The file 643159191Skris crypto/x509v3/ext_dat.h now has the info: this file needs to be 643259191Skris updated whenever a new extension is added to the core code and kept 643359191Skris in ext_nid order. There is a simple program 'tabtest.c' which checks 643459191Skris this. New extensions are not added too often so this file can readily 643559191Skris be maintained manually. 643659191Skris 643759191Skris There are two big advantages in doing things this way. The extensions 643859191Skris can be looked up immediately and no longer need to be "added" using 643959191Skris X509V3_add_standard_extensions(): this function now does nothing. 644059191Skris [Side note: I get *lots* of email saying the extension code doesn't 644159191Skris work because people forget to call this function] 644259191Skris Also no dynamic allocation is done unless new extensions are added: 644359191Skris so if we don't add custom extensions there is no need to call 644459191Skris X509V3_EXT_cleanup(). 644559191Skris [Steve Henson] 644659191Skris 644759191Skris *) Modify enc utility's salting as follows: make salting the default. Add a 644859191Skris magic header, so unsalted files fail gracefully instead of just decrypting 644959191Skris to garbage. This is because not salting is a big security hole, so people 645059191Skris should be discouraged from doing it. 645159191Skris [Ben Laurie] 645259191Skris 645359191Skris *) Fixes and enhancements to the 'x509' utility. It allowed a message 645459191Skris digest to be passed on the command line but it only used this 645559191Skris parameter when signing a certificate. Modified so all relevant 645659191Skris operations are affected by the digest parameter including the 645759191Skris -fingerprint and -x509toreq options. Also -x509toreq choked if a 645859191Skris DSA key was used because it didn't fix the digest. 645959191Skris [Steve Henson] 646059191Skris 646159191Skris *) Initial certificate chain verify code. Currently tests the untrusted 646259191Skris certificates for consistency with the verify purpose (which is set 646359191Skris when the X509_STORE_CTX structure is set up) and checks the pathlength. 646459191Skris 646559191Skris There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 646659191Skris this is because it will reject chains with invalid extensions whereas 646759191Skris every previous version of OpenSSL and SSLeay made no checks at all. 646859191Skris 646959191Skris Trust code: checks the root CA for the relevant trust settings. Trust 647059191Skris settings have an initial value consistent with the verify purpose: e.g. 647159191Skris if the verify purpose is for SSL client use it expects the CA to be 647259191Skris trusted for SSL client use. However the default value can be changed to 647359191Skris permit custom trust settings: one example of this would be to only trust 647459191Skris certificates from a specific "secure" set of CAs. 647559191Skris 647659191Skris Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 647759191Skris which should be used for version portability: especially since the 647859191Skris verify structure is likely to change more often now. 647959191Skris 648059191Skris SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 648159191Skris to set them. If not set then assume SSL clients will verify SSL servers 648259191Skris and vice versa. 648359191Skris 648459191Skris Two new options to the verify program: -untrusted allows a set of 648559191Skris untrusted certificates to be passed in and -purpose which sets the 648659191Skris intended purpose of the certificate. If a purpose is set then the 648759191Skris new chain verify code is used to check extension consistency. 648859191Skris [Steve Henson] 648959191Skris 649059191Skris *) Support for the authority information access extension. 649159191Skris [Steve Henson] 649259191Skris 649359191Skris *) Modify RSA and DSA PEM read routines to transparently handle 649459191Skris PKCS#8 format private keys. New *_PUBKEY_* functions that handle 649559191Skris public keys in a format compatible with certificate 649659191Skris SubjectPublicKeyInfo structures. Unfortunately there were already 649759191Skris functions called *_PublicKey_* which used various odd formats so 649859191Skris these are retained for compatibility: however the DSA variants were 649959191Skris never in a public release so they have been deleted. Changed dsa/rsa 650059191Skris utilities to handle the new format: note no releases ever handled public 650159191Skris keys so we should be OK. 650259191Skris 650359191Skris The primary motivation for this change is to avoid the same fiasco 650459191Skris that dogs private keys: there are several incompatible private key 650559191Skris formats some of which are standard and some OpenSSL specific and 650659191Skris require various evil hacks to allow partial transparent handling and 650759191Skris even then it doesn't work with DER formats. Given the option anything 650859191Skris other than PKCS#8 should be dumped: but the other formats have to 650959191Skris stay in the name of compatibility. 651059191Skris 651159191Skris With public keys and the benefit of hindsight one standard format 651259191Skris is used which works with EVP_PKEY, RSA or DSA structures: though 651359191Skris it clearly returns an error if you try to read the wrong kind of key. 651459191Skris 651559191Skris Added a -pubkey option to the 'x509' utility to output the public key. 651659191Skris Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 651759191Skris (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 651859191Skris EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 651959191Skris that do the same as the EVP_PKEY_assign_*() except they up the 652059191Skris reference count of the added key (they don't "swallow" the 652159191Skris supplied key). 652259191Skris [Steve Henson] 652359191Skris 652459191Skris *) Fixes to crypto/x509/by_file.c the code to read in certificates and 652559191Skris CRLs would fail if the file contained no certificates or no CRLs: 652659191Skris added a new function to read in both types and return the number 652759191Skris read: this means that if none are read it will be an error. The 652859191Skris DER versions of the certificate and CRL reader would always fail 652959191Skris because it isn't possible to mix certificates and CRLs in DER format 653059191Skris without choking one or the other routine. Changed this to just read 653159191Skris a certificate: this is the best we can do. Also modified the code 653259191Skris in apps/verify.c to take notice of return codes: it was previously 653359191Skris attempting to read in certificates from NULL pointers and ignoring 653459191Skris any errors: this is one reason why the cert and CRL reader seemed 653559191Skris to work. It doesn't check return codes from the default certificate 653659191Skris routines: these may well fail if the certificates aren't installed. 653759191Skris [Steve Henson] 653859191Skris 653959191Skris *) Code to support otherName option in GeneralName. 654059191Skris [Steve Henson] 654159191Skris 654259191Skris *) First update to verify code. Change the verify utility 654359191Skris so it warns if it is passed a self signed certificate: 654459191Skris for consistency with the normal behaviour. X509_verify 654559191Skris has been modified to it will now verify a self signed 654659191Skris certificate if *exactly* the same certificate appears 654759191Skris in the store: it was previously impossible to trust a 654859191Skris single self signed certificate. This means that: 654959191Skris openssl verify ss.pem 655059191Skris now gives a warning about a self signed certificate but 655159191Skris openssl verify -CAfile ss.pem ss.pem 655259191Skris is OK. 655359191Skris [Steve Henson] 655459191Skris 655559191Skris *) For servers, store verify_result in SSL_SESSION data structure 655659191Skris (and add it to external session representation). 655759191Skris This is needed when client certificate verifications fails, 655859191Skris but an application-provided verification callback (set by 655959191Skris SSL_CTX_set_cert_verify_callback) allows accepting the session 656059191Skris anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 656159191Skris but returns 1): When the session is reused, we have to set 656259191Skris ssl->verify_result to the appropriate error code to avoid 656359191Skris security holes. 656459191Skris [Bodo Moeller, problem pointed out by Lutz Jaenicke] 656559191Skris 656659191Skris *) Fix a bug in the new PKCS#7 code: it didn't consider the 656759191Skris case in PKCS7_dataInit() where the signed PKCS7 structure 656859191Skris didn't contain any existing data because it was being created. 656959191Skris [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 657059191Skris 657159191Skris *) Add a salt to the key derivation routines in enc.c. This 657259191Skris forms the first 8 bytes of the encrypted file. Also add a 657359191Skris -S option to allow a salt to be input on the command line. 657459191Skris [Steve Henson] 657559191Skris 657659191Skris *) New function X509_cmp(). Oddly enough there wasn't a function 657759191Skris to compare two certificates. We do this by working out the SHA1 657859191Skris hash and comparing that. X509_cmp() will be needed by the trust 657959191Skris code. 658059191Skris [Steve Henson] 658159191Skris 658259191Skris *) SSL_get1_session() is like SSL_get_session(), but increments 658359191Skris the reference count in the SSL_SESSION returned. 658459191Skris [Geoff Thorpe <geoff@eu.c2.net>] 658559191Skris 658659191Skris *) Fix for 'req': it was adding a null to request attributes. 658759191Skris Also change the X509_LOOKUP and X509_INFO code to handle 658859191Skris certificate auxiliary information. 658959191Skris [Steve Henson] 659059191Skris 659159191Skris *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 659259191Skris the 'enc' command. 659359191Skris [Steve Henson] 659459191Skris 659559191Skris *) Add the possibility to add extra information to the memory leak 659659191Skris detecting output, to form tracebacks, showing from where each 659759191Skris allocation was originated: CRYPTO_push_info("constant string") adds 659859191Skris the string plus current file name and line number to a per-thread 659959191Skris stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 660059191Skris is like calling CYRPTO_pop_info() until the stack is empty. 660159191Skris Also updated memory leak detection code to be multi-thread-safe. 660259191Skris [Richard Levitte] 660359191Skris 660459191Skris *) Add options -text and -noout to pkcs7 utility and delete the 660559191Skris encryption options which never did anything. Update docs. 660659191Skris [Steve Henson] 660759191Skris 660859191Skris *) Add options to some of the utilities to allow the pass phrase 660959191Skris to be included on either the command line (not recommended on 661059191Skris OSes like Unix) or read from the environment. Update the 661159191Skris manpages and fix a few bugs. 661259191Skris [Steve Henson] 661359191Skris 661459191Skris *) Add a few manpages for some of the openssl commands. 661559191Skris [Steve Henson] 661659191Skris 661759191Skris *) Fix the -revoke option in ca. It was freeing up memory twice, 661859191Skris leaking and not finding already revoked certificates. 661959191Skris [Steve Henson] 662059191Skris 662159191Skris *) Extensive changes to support certificate auxiliary information. 662259191Skris This involves the use of X509_CERT_AUX structure and X509_AUX 662359191Skris functions. An X509_AUX function such as PEM_read_X509_AUX() 662459191Skris can still read in a certificate file in the usual way but it 662559191Skris will also read in any additional "auxiliary information". By 662659191Skris doing things this way a fair degree of compatibility can be 662759191Skris retained: existing certificates can have this information added 662859191Skris using the new 'x509' options. 662959191Skris 663059191Skris Current auxiliary information includes an "alias" and some trust 663159191Skris settings. The trust settings will ultimately be used in enhanced 663259191Skris certificate chain verification routines: currently a certificate 663359191Skris can only be trusted if it is self signed and then it is trusted 663459191Skris for all purposes. 663559191Skris [Steve Henson] 663659191Skris 663759191Skris *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 663859191Skris The problem was that one of the replacement routines had not been working 663959191Skris since SSLeay releases. For now the offending routine has been replaced 664059191Skris with non-optimised assembler. Even so, this now gives around 95% 664159191Skris performance improvement for 1024 bit RSA signs. 664259191Skris [Mark Cox] 664359191Skris 664459191Skris *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 664559191Skris handling. Most clients have the effective key size in bits equal to 664659191Skris the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 664759191Skris A few however don't do this and instead use the size of the decrypted key 664859191Skris to determine the RC2 key length and the AlgorithmIdentifier to determine 664959191Skris the effective key length. In this case the effective key length can still 665059191Skris be 40 bits but the key length can be 168 bits for example. This is fixed 665159191Skris by manually forcing an RC2 key into the EVP_PKEY structure because the 665259191Skris EVP code can't currently handle unusual RC2 key sizes: it always assumes 665359191Skris the key length and effective key length are equal. 665459191Skris [Steve Henson] 665559191Skris 665659191Skris *) Add a bunch of functions that should simplify the creation of 665759191Skris X509_NAME structures. Now you should be able to do: 665859191Skris X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 665959191Skris and have it automatically work out the correct field type and fill in 666059191Skris the structures. The more adventurous can try: 666159191Skris X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 666259191Skris and it will (hopefully) work out the correct multibyte encoding. 666359191Skris [Steve Henson] 666459191Skris 666559191Skris *) Change the 'req' utility to use the new field handling and multibyte 666659191Skris copy routines. Before the DN field creation was handled in an ad hoc 666759191Skris way in req, ca, and x509 which was rather broken and didn't support 666859191Skris BMPStrings or UTF8Strings. Since some software doesn't implement 666959191Skris BMPStrings or UTF8Strings yet, they can be enabled using the config file 667059191Skris using the dirstring_type option. See the new comment in the default 667159191Skris openssl.cnf for more info. 667259191Skris [Steve Henson] 667359191Skris 667459191Skris *) Make crypto/rand/md_rand.c more robust: 667559191Skris - Assure unique random numbers after fork(). 667659191Skris - Make sure that concurrent threads access the global counter and 667759191Skris md serializably so that we never lose entropy in them 667859191Skris or use exactly the same state in multiple threads. 667959191Skris Access to the large state is not always serializable because 668059191Skris the additional locking could be a performance killer, and 668159191Skris md should be large enough anyway. 668259191Skris [Bodo Moeller] 668359191Skris 668459191Skris *) New file apps/app_rand.c with commonly needed functionality 668559191Skris for handling the random seed file. 668659191Skris 668759191Skris Use the random seed file in some applications that previously did not: 668859191Skris ca, 668959191Skris dsaparam -genkey (which also ignored its '-rand' option), 669059191Skris s_client, 669159191Skris s_server, 669259191Skris x509 (when signing). 669359191Skris Except on systems with /dev/urandom, it is crucial to have a random 669459191Skris seed file at least for key creation, DSA signing, and for DH exchanges; 669559191Skris for RSA signatures we could do without one. 669659191Skris 669759191Skris gendh and gendsa (unlike genrsa) used to read only the first byte 669859191Skris of each file listed in the '-rand' option. The function as previously 669959191Skris found in genrsa is now in app_rand.c and is used by all programs 670059191Skris that support '-rand'. 670159191Skris [Bodo Moeller] 670259191Skris 670359191Skris *) In RAND_write_file, use mode 0600 for creating files; 670459191Skris don't just chmod when it may be too late. 670559191Skris [Bodo Moeller] 670659191Skris 670759191Skris *) Report an error from X509_STORE_load_locations 670859191Skris when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 670959191Skris [Bill Perry] 671059191Skris 671159191Skris *) New function ASN1_mbstring_copy() this copies a string in either 671259191Skris ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 671359191Skris into an ASN1_STRING type. A mask of permissible types is passed 671459191Skris and it chooses the "minimal" type to use or an error if not type 671559191Skris is suitable. 671659191Skris [Steve Henson] 671759191Skris 671859191Skris *) Add function equivalents to the various macros in asn1.h. The old 671959191Skris macros are retained with an M_ prefix. Code inside the library can 672059191Skris use the M_ macros. External code (including the openssl utility) 672159191Skris should *NOT* in order to be "shared library friendly". 672259191Skris [Steve Henson] 672359191Skris 672459191Skris *) Add various functions that can check a certificate's extensions 672559191Skris to see if it usable for various purposes such as SSL client, 672659191Skris server or S/MIME and CAs of these types. This is currently 672759191Skris VERY EXPERIMENTAL but will ultimately be used for certificate chain 672859191Skris verification. Also added a -purpose flag to x509 utility to 672959191Skris print out all the purposes. 673059191Skris [Steve Henson] 673159191Skris 673259191Skris *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 673359191Skris functions. 673459191Skris [Steve Henson] 673559191Skris 673659191Skris *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 673759191Skris for, obtain and decode and extension and obtain its critical flag. 673859191Skris This allows all the necessary extension code to be handled in a 673959191Skris single function call. 674059191Skris [Steve Henson] 674159191Skris 674259191Skris *) RC4 tune-up featuring 30-40% performance improvement on most RISC 674359191Skris platforms. See crypto/rc4/rc4_enc.c for further details. 674459191Skris [Andy Polyakov] 674559191Skris 674659191Skris *) New -noout option to asn1parse. This causes no output to be produced 674759191Skris its main use is when combined with -strparse and -out to extract data 674859191Skris from a file (which may not be in ASN.1 format). 674959191Skris [Steve Henson] 675059191Skris 675159191Skris *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 675259191Skris when producing the local key id. 675359191Skris [Richard Levitte <levitte@stacken.kth.se>] 675459191Skris 675559191Skris *) New option -dhparam in s_server. This allows a DH parameter file to be 675659191Skris stated explicitly. If it is not stated then it tries the first server 675759191Skris certificate file. The previous behaviour hard coded the filename 675859191Skris "server.pem". 675959191Skris [Steve Henson] 676059191Skris 676159191Skris *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 676259191Skris a public key to be input or output. For example: 676359191Skris openssl rsa -in key.pem -pubout -out pubkey.pem 676459191Skris Also added necessary DSA public key functions to handle this. 676559191Skris [Steve Henson] 676659191Skris 676759191Skris *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 676859191Skris in the message. This was handled by allowing 676959191Skris X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 677059191Skris [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 677159191Skris 677259191Skris *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 677359191Skris to the end of the strings whereas this didn't. This would cause problems 677459191Skris if strings read with d2i_ASN1_bytes() were later modified. 677559191Skris [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 677659191Skris 677759191Skris *) Fix for base64 decode bug. When a base64 bio reads only one line of 677859191Skris data and it contains EOF it will end up returning an error. This is 677959191Skris caused by input 46 bytes long. The cause is due to the way base64 678059191Skris BIOs find the start of base64 encoded data. They do this by trying a 678159191Skris trial decode on each line until they find one that works. When they 678259191Skris do a flag is set and it starts again knowing it can pass all the 678359191Skris data directly through the decoder. Unfortunately it doesn't reset 678459191Skris the context it uses. This means that if EOF is reached an attempt 678559191Skris is made to pass two EOFs through the context and this causes the 678659191Skris resulting error. This can also cause other problems as well. As is 678759191Skris usual with these problems it takes *ages* to find and the fix is 678859191Skris trivial: move one line. 678959191Skris [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 679059191Skris 679159191Skris *) Ugly workaround to get s_client and s_server working under Windows. The 679259191Skris old code wouldn't work because it needed to select() on sockets and the 679359191Skris tty (for keypresses and to see if data could be written). Win32 only 679459191Skris supports select() on sockets so we select() with a 1s timeout on the 679559191Skris sockets and then see if any characters are waiting to be read, if none 679659191Skris are present then we retry, we also assume we can always write data to 679759191Skris the tty. This isn't nice because the code then blocks until we've 679859191Skris received a complete line of data and it is effectively polling the 679959191Skris keyboard at 1s intervals: however it's quite a bit better than not 680059191Skris working at all :-) A dedicated Windows application might handle this 680159191Skris with an event loop for example. 680259191Skris [Steve Henson] 680359191Skris 680459191Skris *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 680559191Skris and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 680659191Skris will be called when RSA_sign() and RSA_verify() are used. This is useful 680759191Skris if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 680859191Skris For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 680959191Skris should *not* be used: RSA_sign() and RSA_verify() must be used instead. 681059191Skris This necessitated the support of an extra signature type NID_md5_sha1 681159191Skris for SSL signatures and modifications to the SSL library to use it instead 681259191Skris of calling RSA_public_decrypt() and RSA_private_encrypt(). 681359191Skris [Steve Henson] 681459191Skris 681559191Skris *) Add new -verify -CAfile and -CApath options to the crl program, these 681659191Skris will lookup a CRL issuers certificate and verify the signature in a 681759191Skris similar way to the verify program. Tidy up the crl program so it 681859191Skris no longer accesses structures directly. Make the ASN1 CRL parsing a bit 681959191Skris less strict. It will now permit CRL extensions even if it is not 682059191Skris a V2 CRL: this will allow it to tolerate some broken CRLs. 682159191Skris [Steve Henson] 682259191Skris 682359191Skris *) Initialize all non-automatic variables each time one of the openssl 682459191Skris sub-programs is started (this is necessary as they may be started 682559191Skris multiple times from the "OpenSSL>" prompt). 682659191Skris [Lennart Bang, Bodo Moeller] 682759191Skris 682859191Skris *) Preliminary compilation option RSA_NULL which disables RSA crypto without 682959191Skris removing all other RSA functionality (this is what NO_RSA does). This 683059191Skris is so (for example) those in the US can disable those operations covered 683159191Skris by the RSA patent while allowing storage and parsing of RSA keys and RSA 683259191Skris key generation. 683359191Skris [Steve Henson] 683459191Skris 683559191Skris *) Non-copying interface to BIO pairs. 683659191Skris (still largely untested) 683759191Skris [Bodo Moeller] 683859191Skris 683959191Skris *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 684059191Skris ASCII string. This was handled independently in various places before. 684159191Skris [Steve Henson] 684259191Skris 684359191Skris *) New functions UTF8_getc() and UTF8_putc() that parse and generate 684459191Skris UTF8 strings a character at a time. 684559191Skris [Steve Henson] 684659191Skris 684759191Skris *) Use client_version from client hello to select the protocol 684859191Skris (s23_srvr.c) and for RSA client key exchange verification 684959191Skris (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 685059191Skris [Bodo Moeller] 685159191Skris 685259191Skris *) Add various utility functions to handle SPKACs, these were previously 685359191Skris handled by poking round in the structure internals. Added new function 685459191Skris NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 685559191Skris print, verify and generate SPKACs. Based on an original idea from 685659191Skris Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 685759191Skris [Steve Henson] 685859191Skris 685959191Skris *) RIPEMD160 is operational on all platforms and is back in 'make test'. 686059191Skris [Andy Polyakov] 686159191Skris 686259191Skris *) Allow the config file extension section to be overwritten on the 686359191Skris command line. Based on an original idea from Massimiliano Pala 686459191Skris <madwolf@comune.modena.it>. The new option is called -extensions 686559191Skris and can be applied to ca, req and x509. Also -reqexts to override 686659191Skris the request extensions in req and -crlexts to override the crl extensions 686759191Skris in ca. 686859191Skris [Steve Henson] 686959191Skris 687059191Skris *) Add new feature to the SPKAC handling in ca. Now you can include 687159191Skris the same field multiple times by preceding it by "XXXX." for example: 687259191Skris 1.OU="Unit name 1" 687359191Skris 2.OU="Unit name 2" 687459191Skris this is the same syntax as used in the req config file. 687559191Skris [Steve Henson] 687659191Skris 687759191Skris *) Allow certificate extensions to be added to certificate requests. These 687859191Skris are specified in a 'req_extensions' option of the req section of the 687959191Skris config file. They can be printed out with the -text option to req but 688059191Skris are otherwise ignored at present. 688159191Skris [Steve Henson] 688259191Skris 688359191Skris *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 688459191Skris data read consists of only the final block it would not decrypted because 688559191Skris EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 688659191Skris A misplaced 'break' also meant the decrypted final block might not be 688759191Skris copied until the next read. 688859191Skris [Steve Henson] 688959191Skris 689059191Skris *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 689159191Skris a few extra parameters to the DH structure: these will be useful if 689259191Skris for example we want the value of 'q' or implement X9.42 DH. 689359191Skris [Steve Henson] 689459191Skris 689559191Skris *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 689659191Skris provides hooks that allow the default DSA functions or functions on a 689759191Skris "per key" basis to be replaced. This allows hardware acceleration and 689859191Skris hardware key storage to be handled without major modification to the 689959191Skris library. Also added low level modexp hooks and CRYPTO_EX structure and 690059191Skris associated functions. 690159191Skris [Steve Henson] 690259191Skris 690359191Skris *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 690459191Skris as "read only": it can't be written to and the buffer it points to will 690559191Skris not be freed. Reading from a read only BIO is much more efficient than 690659191Skris a normal memory BIO. This was added because there are several times when 690759191Skris an area of memory needs to be read from a BIO. The previous method was 690859191Skris to create a memory BIO and write the data to it, this results in two 690959191Skris copies of the data and an O(n^2) reading algorithm. There is a new 691059191Skris function BIO_new_mem_buf() which creates a read only memory BIO from 691159191Skris an area of memory. Also modified the PKCS#7 routines to use read only 691259191Skris memory BIOs. 691359191Skris [Steve Henson] 691459191Skris 691559191Skris *) Bugfix: ssl23_get_client_hello did not work properly when called in 691659191Skris state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 691759191Skris a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 691859191Skris but a retry condition occured while trying to read the rest. 691959191Skris [Bodo Moeller] 692059191Skris 692159191Skris *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 692259191Skris NID_pkcs7_encrypted by default: this was wrong since this should almost 692359191Skris always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 692459191Skris the encrypted data type: this is a more sensible place to put it and it 692559191Skris allows the PKCS#12 code to be tidied up that duplicated this 692659191Skris functionality. 692759191Skris [Steve Henson] 692859191Skris 692959191Skris *) Changed obj_dat.pl script so it takes its input and output files on 693059191Skris the command line. This should avoid shell escape redirection problems 693159191Skris under Win32. 693259191Skris [Steve Henson] 693359191Skris 693459191Skris *) Initial support for certificate extension requests, these are included 693559191Skris in things like Xenroll certificate requests. Included functions to allow 693659191Skris extensions to be obtained and added. 693759191Skris [Steve Henson] 693859191Skris 693959191Skris *) -crlf option to s_client and s_server for sending newlines as 694059191Skris CRLF (as required by many protocols). 694159191Skris [Bodo Moeller] 694259191Skris 694355714Skris Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 694455714Skris 694555714Skris *) Install libRSAglue.a when OpenSSL is built with RSAref. 694655714Skris [Ralf S. Engelschall] 694755714Skris 694855714Skris *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 694955714Skris [Andrija Antonijevic <TheAntony2@bigfoot.com>] 695055714Skris 695155714Skris *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 695255714Skris program. 695355714Skris [Steve Henson] 695455714Skris 695555714Skris *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 695655714Skris DH parameters/keys (q is lost during that conversion, but the resulting 695755714Skris DH parameters contain its length). 695855714Skris 695955714Skris For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 696055714Skris much faster than DH_generate_parameters (which creates parameters 696155714Skris where p = 2*q + 1), and also the smaller q makes DH computations 696255714Skris much more efficient (160-bit exponentiation instead of 1024-bit 696355714Skris exponentiation); so this provides a convenient way to support DHE 696455714Skris ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 696555714Skris utter importance to use 696655714Skris SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 696755714Skris or 696855714Skris SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 696955714Skris when such DH parameters are used, because otherwise small subgroup 697055714Skris attacks may become possible! 697155714Skris [Bodo Moeller] 697255714Skris 697355714Skris *) Avoid memory leak in i2d_DHparams. 697455714Skris [Bodo Moeller] 697555714Skris 697655714Skris *) Allow the -k option to be used more than once in the enc program: 697755714Skris this allows the same encrypted message to be read by multiple recipients. 697855714Skris [Steve Henson] 697955714Skris 698055714Skris *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 698155714Skris an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 698255714Skris it will always use the numerical form of the OID, even if it has a short 698355714Skris or long name. 698455714Skris [Steve Henson] 698555714Skris 698655714Skris *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 698755714Skris method only got called if p,q,dmp1,dmq1,iqmp components were present, 698855714Skris otherwise bn_mod_exp was called. In the case of hardware keys for example 698955714Skris no private key components need be present and it might store extra data 699059191Skris in the RSA structure, which cannot be accessed from bn_mod_exp. 699159191Skris By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 699259191Skris private key operations. 699355714Skris [Steve Henson] 699455714Skris 699555714Skris *) Added support for SPARC Linux. 699655714Skris [Andy Polyakov] 699755714Skris 699855714Skris *) pem_password_cb function type incompatibly changed from 699955714Skris typedef int pem_password_cb(char *buf, int size, int rwflag); 700055714Skris to 700155714Skris ....(char *buf, int size, int rwflag, void *userdata); 700255714Skris so that applications can pass data to their callbacks: 700355714Skris The PEM[_ASN1]_{read,write}... functions and macros now take an 700455714Skris additional void * argument, which is just handed through whenever 700555714Skris the password callback is called. 700659191Skris [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 700755714Skris 700855714Skris New function SSL_CTX_set_default_passwd_cb_userdata. 700955714Skris 701055714Skris Compatibility note: As many C implementations push function arguments 701155714Skris onto the stack in reverse order, the new library version is likely to 701255714Skris interoperate with programs that have been compiled with the old 701355714Skris pem_password_cb definition (PEM_whatever takes some data that 701455714Skris happens to be on the stack as its last argument, and the callback 701555714Skris just ignores this garbage); but there is no guarantee whatsoever that 701655714Skris this will work. 701755714Skris 701855714Skris *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 701955714Skris (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 702055714Skris problems not only on Windows, but also on some Unix platforms. 702155714Skris To avoid problematic command lines, these definitions are now in an 702255714Skris auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 702355714Skris for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 702455714Skris [Bodo Moeller] 702555714Skris 702655714Skris *) MIPS III/IV assembler module is reimplemented. 702755714Skris [Andy Polyakov] 702855714Skris 702955714Skris *) More DES library cleanups: remove references to srand/rand and 703055714Skris delete an unused file. 703155714Skris [Ulf M�ller] 703255714Skris 703355714Skris *) Add support for the the free Netwide assembler (NASM) under Win32, 703455714Skris since not many people have MASM (ml) and it can be hard to obtain. 703555714Skris This is currently experimental but it seems to work OK and pass all 703655714Skris the tests. Check out INSTALL.W32 for info. 703755714Skris [Steve Henson] 703855714Skris 703955714Skris *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 704055714Skris without temporary keys kept an extra copy of the server key, 704155714Skris and connections with temporary keys did not free everything in case 704255714Skris of an error. 704355714Skris [Bodo Moeller] 704455714Skris 704555714Skris *) New function RSA_check_key and new openssl rsa option -check 704655714Skris for verifying the consistency of RSA keys. 704755714Skris [Ulf Moeller, Bodo Moeller] 704855714Skris 704955714Skris *) Various changes to make Win32 compile work: 705055714Skris 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 705155714Skris 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 705255714Skris comparison" warnings. 705355714Skris 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 705455714Skris [Steve Henson] 705555714Skris 705655714Skris *) Add a debugging option to PKCS#5 v2 key generation function: when 705755714Skris you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 705855714Skris derived keys are printed to stderr. 705955714Skris [Steve Henson] 706055714Skris 706155714Skris *) Copy the flags in ASN1_STRING_dup(). 706255714Skris [Roman E. Pavlov <pre@mo.msk.ru>] 706355714Skris 706455714Skris *) The x509 application mishandled signing requests containing DSA 706555714Skris keys when the signing key was also DSA and the parameters didn't match. 706655714Skris 706755714Skris It was supposed to omit the parameters when they matched the signing key: 706855714Skris the verifying software was then supposed to automatically use the CA's 706955714Skris parameters if they were absent from the end user certificate. 707055714Skris 707155714Skris Omitting parameters is no longer recommended. The test was also 707255714Skris the wrong way round! This was probably due to unusual behaviour in 707355714Skris EVP_cmp_parameters() which returns 1 if the parameters match. 707455714Skris This meant that parameters were omitted when they *didn't* match and 707555714Skris the certificate was useless. Certificates signed with 'ca' didn't have 707655714Skris this bug. 707755714Skris [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 707855714Skris 707955714Skris *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 708055714Skris The interface is as follows: 708155714Skris Applications can use 708255714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 708355714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 708455714Skris "off" is now the default. 708555714Skris The library internally uses 708655714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 708755714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 708855714Skris to disable memory-checking temporarily. 708955714Skris 709055714Skris Some inconsistent states that previously were possible (and were 709155714Skris even the default) are now avoided. 709255714Skris 709355714Skris -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 709455714Skris with each memory chunk allocated; this is occasionally more helpful 709555714Skris than just having a counter. 709655714Skris 709755714Skris -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 709855714Skris 709955714Skris -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 710055714Skris extensions. 710155714Skris [Bodo Moeller] 710255714Skris 710355714Skris *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 710455714Skris which largely parallels "options", but is for changing API behaviour, 710555714Skris whereas "options" are about protocol behaviour. 710655714Skris Initial "mode" flags are: 710755714Skris 710855714Skris SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 710955714Skris a single record has been written. 711055714Skris SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 711155714Skris retries use the same buffer location. 711255714Skris (But all of the contents must be 711355714Skris copied!) 711455714Skris [Bodo Moeller] 711555714Skris 711679998Skris *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 711755714Skris worked. 711855714Skris 711955714Skris *) Fix problems with no-hmac etc. 712055714Skris [Ulf M�ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 712155714Skris 712255714Skris *) New functions RSA_get_default_method(), RSA_set_method() and 712355714Skris RSA_get_method(). These allows replacement of RSA_METHODs without having 712455714Skris to mess around with the internals of an RSA structure. 712555714Skris [Steve Henson] 712655714Skris 712755714Skris *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 712855714Skris Also really enable memory leak checks in openssl.c and in some 712955714Skris test programs. 713055714Skris [Chad C. Mulligan, Bodo Moeller] 713155714Skris 713255714Skris *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 713355714Skris up the length of negative integers. This has now been simplified to just 713455714Skris store the length when it is first determined and use it later, rather 713555714Skris than trying to keep track of where data is copied and updating it to 713655714Skris point to the end. 713755714Skris [Steve Henson, reported by Brien Wheeler 713855714Skris <bwheeler@authentica-security.com>] 713955714Skris 714055714Skris *) Add a new function PKCS7_signatureVerify. This allows the verification 714155714Skris of a PKCS#7 signature but with the signing certificate passed to the 714255714Skris function itself. This contrasts with PKCS7_dataVerify which assumes the 714355714Skris certificate is present in the PKCS#7 structure. This isn't always the 714455714Skris case: certificates can be omitted from a PKCS#7 structure and be 714555714Skris distributed by "out of band" means (such as a certificate database). 714655714Skris [Steve Henson] 714755714Skris 714855714Skris *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 714955714Skris function prototypes in pem.h, also change util/mkdef.pl to add the 715055714Skris necessary function names. 715155714Skris [Steve Henson] 715255714Skris 715355714Skris *) mk1mf.pl (used by Windows builds) did not properly read the 715455714Skris options set by Configure in the top level Makefile, and Configure 715555714Skris was not even able to write more than one option correctly. 715655714Skris Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 715755714Skris [Bodo Moeller] 715855714Skris 715955714Skris *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 716055714Skris file to be loaded from a BIO or FILE pointer. The BIO version will 716155714Skris for example allow memory BIOs to contain config info. 716255714Skris [Steve Henson] 716355714Skris 716455714Skris *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 716555714Skris Whoever hopes to achieve shared-library compatibility across versions 716655714Skris must use this, not the compile-time macro. 716755714Skris (Exercise 0.9.4: Which is the minimum library version required by 716855714Skris such programs?) 716955714Skris Note: All this applies only to multi-threaded programs, others don't 717055714Skris need locks. 717155714Skris [Bodo Moeller] 717255714Skris 717355714Skris *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 717455714Skris through a BIO pair triggered the default case, i.e. 717555714Skris SSLerr(...,SSL_R_UNKNOWN_STATE). 717655714Skris [Bodo Moeller] 717755714Skris 717855714Skris *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 717955714Skris can use the SSL library even if none of the specific BIOs is 718055714Skris appropriate. 718155714Skris [Bodo Moeller] 718255714Skris 718355714Skris *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 718455714Skris for the encoded length. 718555714Skris [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 718655714Skris 718755714Skris *) Add initial documentation of the X509V3 functions. 718855714Skris [Steve Henson] 718955714Skris 719055714Skris *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 719155714Skris PEM_write_bio_PKCS8PrivateKey() that are equivalent to 719255714Skris PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 719355714Skris secure PKCS#8 private key format with a high iteration count. 719455714Skris [Steve Henson] 719555714Skris 719655714Skris *) Fix determination of Perl interpreter: A perl or perl5 719755714Skris _directory_ in $PATH was also accepted as the interpreter. 719855714Skris [Ralf S. Engelschall] 719955714Skris 720055714Skris *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 720155714Skris wrong with it but it was very old and did things like calling 720255714Skris PEM_ASN1_read() directly and used MD5 for the hash not to mention some 720355714Skris unusual formatting. 720455714Skris [Steve Henson] 720555714Skris 720655714Skris *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 720755714Skris to use the new extension code. 720855714Skris [Steve Henson] 720955714Skris 721055714Skris *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 721155714Skris with macros. This should make it easier to change their form, add extra 721255714Skris arguments etc. Fix a few PEM prototypes which didn't have cipher as a 721355714Skris constant. 721455714Skris [Steve Henson] 721555714Skris 721655714Skris *) Add to configuration table a new entry that can specify an alternative 721755714Skris name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 721855714Skris according to Mark Crispin <MRC@Panda.COM>. 721955714Skris [Bodo Moeller] 722055714Skris 722155714Skris#if 0 722255714Skris *) DES CBC did not update the IV. Weird. 722355714Skris [Ben Laurie] 722455714Skris#else 722555714Skris des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 722655714Skris Changing the behaviour of the former might break existing programs -- 722755714Skris where IV updating is needed, des_ncbc_encrypt can be used. 722855714Skris#endif 722955714Skris 723055714Skris *) When bntest is run from "make test" it drives bc to check its 723155714Skris calculations, as well as internally checking them. If an internal check 723255714Skris fails, it needs to cause bc to give a non-zero result or make test carries 723355714Skris on without noticing the failure. Fixed. 723455714Skris [Ben Laurie] 723555714Skris 723655714Skris *) DES library cleanups. 723755714Skris [Ulf M�ller] 723855714Skris 723955714Skris *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 724055714Skris used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 724155714Skris ciphers. NOTE: although the key derivation function has been verified 724255714Skris against some published test vectors it has not been extensively tested 724355714Skris yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 724455714Skris of v2.0. 724555714Skris [Steve Henson] 724655714Skris 724755714Skris *) Instead of "mkdir -p", which is not fully portable, use new 724855714Skris Perl script "util/mkdir-p.pl". 724955714Skris [Bodo Moeller] 725055714Skris 725155714Skris *) Rewrite the way password based encryption (PBE) is handled. It used to 725255714Skris assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 725355714Skris structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 725455714Skris but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 725555714Skris the 'parameter' field of the AlgorithmIdentifier is passed to the 725655714Skris underlying key generation function so it must do its own ASN1 parsing. 725755714Skris This has also changed the EVP_PBE_CipherInit() function which now has a 725855714Skris 'parameter' argument instead of literal salt and iteration count values 725955714Skris and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 726055714Skris [Steve Henson] 726155714Skris 726255714Skris *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 726355714Skris and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 726455714Skris Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 726555714Skris KEY" because this clashed with PKCS#8 unencrypted string. Since this 726655714Skris value was just used as a "magic string" and not used directly its 726755714Skris value doesn't matter. 726855714Skris [Steve Henson] 726955714Skris 727055714Skris *) Introduce some semblance of const correctness to BN. Shame C doesn't 727155714Skris support mutable. 727255714Skris [Ben Laurie] 727355714Skris 727455714Skris *) "linux-sparc64" configuration (ultrapenguin). 727555714Skris [Ray Miller <ray.miller@oucs.ox.ac.uk>] 727655714Skris "linux-sparc" configuration. 727755714Skris [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 727855714Skris 727955714Skris *) config now generates no-xxx options for missing ciphers. 728055714Skris [Ulf M�ller] 728155714Skris 728255714Skris *) Support the EBCDIC character set (work in progress). 728355714Skris File ebcdic.c not yet included because it has a different license. 728455714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 728555714Skris 728655714Skris *) Support BS2000/OSD-POSIX. 728755714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 728855714Skris 728955714Skris *) Make callbacks for key generation use void * instead of char *. 729055714Skris [Ben Laurie] 729155714Skris 729255714Skris *) Make S/MIME samples compile (not yet tested). 729355714Skris [Ben Laurie] 729455714Skris 729555714Skris *) Additional typesafe stacks. 729655714Skris [Ben Laurie] 729755714Skris 729855714Skris *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 729955714Skris [Bodo Moeller] 730055714Skris 730155714Skris 730255714Skris Changes between 0.9.3 and 0.9.3a [29 May 1999] 730355714Skris 730455714Skris *) New configuration variant "sco5-gcc". 730555714Skris 730655714Skris *) Updated some demos. 730755714Skris [Sean O Riordain, Wade Scholine] 730855714Skris 730955714Skris *) Add missing BIO_free at exit of pkcs12 application. 731055714Skris [Wu Zhigang] 731155714Skris 731255714Skris *) Fix memory leak in conf.c. 731355714Skris [Steve Henson] 731455714Skris 731555714Skris *) Updates for Win32 to assembler version of MD5. 731655714Skris [Steve Henson] 731755714Skris 731855714Skris *) Set #! path to perl in apps/der_chop to where we found it 731955714Skris instead of using a fixed path. 732055714Skris [Bodo Moeller] 732155714Skris 732255714Skris *) SHA library changes for irix64-mips4-cc. 732355714Skris [Andy Polyakov] 732455714Skris 732555714Skris *) Improvements for VMS support. 732655714Skris [Richard Levitte] 732755714Skris 732855714Skris 732955714Skris Changes between 0.9.2b and 0.9.3 [24 May 1999] 733055714Skris 733155714Skris *) Bignum library bug fix. IRIX 6 passes "make test" now! 733255714Skris This also avoids the problems with SC4.2 and unpatched SC5. 733355714Skris [Andy Polyakov <appro@fy.chalmers.se>] 733455714Skris 733555714Skris *) New functions sk_num, sk_value and sk_set to replace the previous macros. 733655714Skris These are required because of the typesafe stack would otherwise break 733755714Skris existing code. If old code used a structure member which used to be STACK 733855714Skris and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 733955714Skris sk_num or sk_value it would produce an error because the num, data members 734055714Skris are not present in STACK_OF. Now it just produces a warning. sk_set 734155714Skris replaces the old method of assigning a value to sk_value 734255714Skris (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 734355714Skris that does this will no longer work (and should use sk_set instead) but 734455714Skris this could be regarded as a "questionable" behaviour anyway. 734555714Skris [Steve Henson] 734655714Skris 734755714Skris *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 734855714Skris correctly handle encrypted S/MIME data. 734955714Skris [Steve Henson] 735055714Skris 735155714Skris *) Change type of various DES function arguments from des_cblock 735255714Skris (which means, in function argument declarations, pointer to char) 735355714Skris to des_cblock * (meaning pointer to array with 8 char elements), 735455714Skris which allows the compiler to do more typechecking; it was like 735555714Skris that back in SSLeay, but with lots of ugly casts. 735655714Skris 735755714Skris Introduce new type const_des_cblock. 735855714Skris [Bodo Moeller] 735955714Skris 736055714Skris *) Reorganise the PKCS#7 library and get rid of some of the more obvious 736155714Skris problems: find RecipientInfo structure that matches recipient certificate 736255714Skris and initialise the ASN1 structures properly based on passed cipher. 736355714Skris [Steve Henson] 736455714Skris 736555714Skris *) Belatedly make the BN tests actually check the results. 736655714Skris [Ben Laurie] 736755714Skris 736855714Skris *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 736955714Skris to and from BNs: it was completely broken. New compilation option 737055714Skris NEG_PUBKEY_BUG to allow for some broken certificates that encode public 737155714Skris key elements as negative integers. 737255714Skris [Steve Henson] 737355714Skris 737455714Skris *) Reorganize and speed up MD5. 737555714Skris [Andy Polyakov <appro@fy.chalmers.se>] 737655714Skris 737755714Skris *) VMS support. 737855714Skris [Richard Levitte <richard@levitte.org>] 737955714Skris 738055714Skris *) New option -out to asn1parse to allow the parsed structure to be 738155714Skris output to a file. This is most useful when combined with the -strparse 738255714Skris option to examine the output of things like OCTET STRINGS. 738355714Skris [Steve Henson] 738455714Skris 738555714Skris *) Make SSL library a little more fool-proof by not requiring any longer 738655714Skris that SSL_set_{accept,connect}_state be called before 738755714Skris SSL_{accept,connect} may be used (SSL_set_..._state is omitted 738855714Skris in many applications because usually everything *appeared* to work as 738955714Skris intended anyway -- now it really works as intended). 739055714Skris [Bodo Moeller] 739155714Skris 739255714Skris *) Move openssl.cnf out of lib/. 739355714Skris [Ulf M�ller] 739455714Skris 739555714Skris *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 739655714Skris -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 739755714Skris -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 739855714Skris [Ralf S. Engelschall] 739955714Skris 740055714Skris *) Various fixes to the EVP and PKCS#7 code. It may now be able to 740155714Skris handle PKCS#7 enveloped data properly. 740255714Skris [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 740355714Skris 740455714Skris *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 740555714Skris copying pointers. The cert_st handling is changed by this in 740655714Skris various ways (and thus what used to be known as ctx->default_cert 740755714Skris is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 740855714Skris any longer when s->cert does not give us what we need). 740955714Skris ssl_cert_instantiate becomes obsolete by this change. 741055714Skris As soon as we've got the new code right (possibly it already is?), 741155714Skris we have solved a couple of bugs of the earlier code where s->cert 741255714Skris was used as if it could not have been shared with other SSL structures. 741355714Skris 741455714Skris Note that using the SSL API in certain dirty ways now will result 741555714Skris in different behaviour than observed with earlier library versions: 741655714Skris Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 741755714Skris does not influence s as it used to. 741855714Skris 741955714Skris In order to clean up things more thoroughly, inside SSL_SESSION 742055714Skris we don't use CERT any longer, but a new structure SESS_CERT 742155714Skris that holds per-session data (if available); currently, this is 742255714Skris the peer's certificate chain and, for clients, the server's certificate 742355714Skris and temporary key. CERT holds only those values that can have 742455714Skris meaningful defaults in an SSL_CTX. 742555714Skris [Bodo Moeller] 742655714Skris 742755714Skris *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 742855714Skris from the internal representation. Various PKCS#7 fixes: remove some 742955714Skris evil casts and set the enc_dig_alg field properly based on the signing 743055714Skris key type. 743155714Skris [Steve Henson] 743255714Skris 743355714Skris *) Allow PKCS#12 password to be set from the command line or the 743455714Skris environment. Let 'ca' get its config file name from the environment 743555714Skris variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 743655714Skris and 'x509'). 743755714Skris [Steve Henson] 743855714Skris 743955714Skris *) Allow certificate policies extension to use an IA5STRING for the 744055714Skris organization field. This is contrary to the PKIX definition but 744155714Skris VeriSign uses it and IE5 only recognises this form. Document 'x509' 744255714Skris extension option. 744355714Skris [Steve Henson] 744455714Skris 744555714Skris *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 744655714Skris without disallowing inline assembler and the like for non-pedantic builds. 744755714Skris [Ben Laurie] 744855714Skris 744955714Skris *) Support Borland C++ builder. 745055714Skris [Janez Jere <jj@void.si>, modified by Ulf M�ller] 745155714Skris 745255714Skris *) Support Mingw32. 745355714Skris [Ulf M�ller] 745455714Skris 745555714Skris *) SHA-1 cleanups and performance enhancements. 745655714Skris [Andy Polyakov <appro@fy.chalmers.se>] 745755714Skris 745855714Skris *) Sparc v8plus assembler for the bignum library. 745955714Skris [Andy Polyakov <appro@fy.chalmers.se>] 746055714Skris 746155714Skris *) Accept any -xxx and +xxx compiler options in Configure. 746255714Skris [Ulf M�ller] 746355714Skris 746455714Skris *) Update HPUX configuration. 746555714Skris [Anonymous] 746655714Skris 746755714Skris *) Add missing sk_<type>_unshift() function to safestack.h 746855714Skris [Ralf S. Engelschall] 746955714Skris 747055714Skris *) New function SSL_CTX_use_certificate_chain_file that sets the 747155714Skris "extra_cert"s in addition to the certificate. (This makes sense 747255714Skris only for "PEM" format files, as chains as a whole are not 747355714Skris DER-encoded.) 747455714Skris [Bodo Moeller] 747555714Skris 747655714Skris *) Support verify_depth from the SSL API. 747755714Skris x509_vfy.c had what can be considered an off-by-one-error: 747855714Skris Its depth (which was not part of the external interface) 747955714Skris was actually counting the number of certificates in a chain; 748055714Skris now it really counts the depth. 748155714Skris [Bodo Moeller] 748255714Skris 748355714Skris *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 748455714Skris instead of X509err, which often resulted in confusing error 748555714Skris messages since the error codes are not globally unique 748655714Skris (e.g. an alleged error in ssl3_accept when a certificate 748755714Skris didn't match the private key). 748855714Skris 748955714Skris *) New function SSL_CTX_set_session_id_context that allows to set a default 749055714Skris value (so that you don't need SSL_set_session_id_context for each 749155714Skris connection using the SSL_CTX). 749255714Skris [Bodo Moeller] 749355714Skris 749455714Skris *) OAEP decoding bug fix. 749555714Skris [Ulf M�ller] 749655714Skris 749755714Skris *) Support INSTALL_PREFIX for package builders, as proposed by 749855714Skris David Harris. 749955714Skris [Bodo Moeller] 750055714Skris 750155714Skris *) New Configure options "threads" and "no-threads". For systems 750255714Skris where the proper compiler options are known (currently Solaris 750355714Skris and Linux), "threads" is the default. 750455714Skris [Bodo Moeller] 750555714Skris 750655714Skris *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 750755714Skris [Bodo Moeller] 750855714Skris 750955714Skris *) Install various scripts to $(OPENSSLDIR)/misc, not to 751055714Skris $(INSTALLTOP)/bin -- they shouldn't clutter directories 751155714Skris such as /usr/local/bin. 751255714Skris [Bodo Moeller] 751355714Skris 751455714Skris *) "make linux-shared" to build shared libraries. 751555714Skris [Niels Poppe <niels@netbox.org>] 751655714Skris 751755714Skris *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 751855714Skris [Ulf M�ller] 751955714Skris 752055714Skris *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 752155714Skris extension adding in x509 utility. 752255714Skris [Steve Henson] 752355714Skris 752455714Skris *) Remove NOPROTO sections and error code comments. 752555714Skris [Ulf M�ller] 752655714Skris 752755714Skris *) Partial rewrite of the DEF file generator to now parse the ANSI 752855714Skris prototypes. 752955714Skris [Steve Henson] 753055714Skris 753155714Skris *) New Configure options --prefix=DIR and --openssldir=DIR. 753255714Skris [Ulf M�ller] 753355714Skris 753455714Skris *) Complete rewrite of the error code script(s). It is all now handled 753555714Skris by one script at the top level which handles error code gathering, 753655714Skris header rewriting and C source file generation. It should be much better 753755714Skris than the old method: it now uses a modified version of Ulf's parser to 753855714Skris read the ANSI prototypes in all header files (thus the old K&R definitions 753955714Skris aren't needed for error creation any more) and do a better job of 754055714Skris translating function codes into names. The old 'ASN1 error code imbedded 754155714Skris in a comment' is no longer necessary and it doesn't use .err files which 754255714Skris have now been deleted. Also the error code call doesn't have to appear all 754355714Skris on one line (which resulted in some large lines...). 754455714Skris [Steve Henson] 754555714Skris 754655714Skris *) Change #include filenames from <foo.h> to <openssl/foo.h>. 754755714Skris [Bodo Moeller] 754855714Skris 754955714Skris *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 755055714Skris 0 (which usually indicates a closed connection), but continue reading. 755155714Skris [Bodo Moeller] 755255714Skris 755355714Skris *) Fix some race conditions. 755455714Skris [Bodo Moeller] 755555714Skris 755655714Skris *) Add support for CRL distribution points extension. Add Certificate 755755714Skris Policies and CRL distribution points documentation. 755855714Skris [Steve Henson] 755955714Skris 756055714Skris *) Move the autogenerated header file parts to crypto/opensslconf.h. 756155714Skris [Ulf M�ller] 756255714Skris 756355714Skris *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 756455714Skris 8 of keying material. Merlin has also confirmed interop with this fix 756555714Skris between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 756655714Skris [Merlin Hughes <merlin@baltimore.ie>] 756755714Skris 756855714Skris *) Fix lots of warnings. 756955714Skris [Richard Levitte <levitte@stacken.kth.se>] 757055714Skris 757155714Skris *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 757255714Skris the directory spec didn't end with a LIST_SEPARATOR_CHAR. 757355714Skris [Richard Levitte <levitte@stacken.kth.se>] 757455714Skris 757555714Skris *) Fix problems with sizeof(long) == 8. 757655714Skris [Andy Polyakov <appro@fy.chalmers.se>] 757755714Skris 757855714Skris *) Change functions to ANSI C. 757955714Skris [Ulf M�ller] 758055714Skris 758155714Skris *) Fix typos in error codes. 758255714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M�ller] 758355714Skris 758455714Skris *) Remove defunct assembler files from Configure. 758555714Skris [Ulf M�ller] 758655714Skris 758755714Skris *) SPARC v8 assembler BIGNUM implementation. 758855714Skris [Andy Polyakov <appro@fy.chalmers.se>] 758955714Skris 759055714Skris *) Support for Certificate Policies extension: both print and set. 759155714Skris Various additions to support the r2i method this uses. 759255714Skris [Steve Henson] 759355714Skris 759455714Skris *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 759555714Skris return a const string when you are expecting an allocated buffer. 759655714Skris [Ben Laurie] 759755714Skris 759855714Skris *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 759955714Skris types DirectoryString and DisplayText. 760055714Skris [Steve Henson] 760155714Skris 760255714Skris *) Add code to allow r2i extensions to access the configuration database, 760355714Skris add an LHASH database driver and add several ctx helper functions. 760455714Skris [Steve Henson] 760555714Skris 760655714Skris *) Fix an evil bug in bn_expand2() which caused various BN functions to 760755714Skris fail when they extended the size of a BIGNUM. 760855714Skris [Steve Henson] 760955714Skris 761055714Skris *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 761155714Skris support typesafe stack. 761255714Skris [Steve Henson] 761355714Skris 761455714Skris *) Fix typo in SSL_[gs]et_options(). 761555714Skris [Nils Frostberg <nils@medcom.se>] 761655714Skris 761755714Skris *) Delete various functions and files that belonged to the (now obsolete) 761855714Skris old X509V3 handling code. 761955714Skris [Steve Henson] 762055714Skris 762155714Skris *) New Configure option "rsaref". 762255714Skris [Ulf M�ller] 762355714Skris 762455714Skris *) Don't auto-generate pem.h. 762555714Skris [Bodo Moeller] 762655714Skris 762755714Skris *) Introduce type-safe ASN.1 SETs. 762855714Skris [Ben Laurie] 762955714Skris 763055714Skris *) Convert various additional casted stacks to type-safe STACK_OF() variants. 763155714Skris [Ben Laurie, Ralf S. Engelschall, Steve Henson] 763255714Skris 763355714Skris *) Introduce type-safe STACKs. This will almost certainly break lots of code 763455714Skris that links with OpenSSL (well at least cause lots of warnings), but fear 763555714Skris not: the conversion is trivial, and it eliminates loads of evil casts. A 763655714Skris few STACKed things have been converted already. Feel free to convert more. 763755714Skris In the fullness of time, I'll do away with the STACK type altogether. 763855714Skris [Ben Laurie] 763955714Skris 764055714Skris *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 764155714Skris specified in <certfile> by updating the entry in the index.txt file. 764255714Skris This way one no longer has to edit the index.txt file manually for 764355714Skris revoking a certificate. The -revoke option does the gory details now. 764455714Skris [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 764555714Skris 764655714Skris *) Fix `openssl crl -noout -text' combination where `-noout' killed the 764755714Skris `-text' option at all and this way the `-noout -text' combination was 764855714Skris inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 764955714Skris [Ralf S. Engelschall] 765055714Skris 765155714Skris *) Make sure a corresponding plain text error message exists for the 765255714Skris X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 765355714Skris verify callback function determined that a certificate was revoked. 765455714Skris [Ralf S. Engelschall] 765555714Skris 765655714Skris *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 765755714Skris ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 765855714Skris all available cipers including rc5, which was forgotten until now. 765955714Skris In order to let the testing shell script know which algorithms 766055714Skris are available, a new (up to now undocumented) command 766155714Skris "openssl list-cipher-commands" is used. 766255714Skris [Bodo Moeller] 766355714Skris 766455714Skris *) Bugfix: s_client occasionally would sleep in select() when 766555714Skris it should have checked SSL_pending() first. 766655714Skris [Bodo Moeller] 766755714Skris 766855714Skris *) New functions DSA_do_sign and DSA_do_verify to provide access to 766955714Skris the raw DSA values prior to ASN.1 encoding. 767055714Skris [Ulf M�ller] 767155714Skris 767255714Skris *) Tweaks to Configure 767355714Skris [Niels Poppe <niels@netbox.org>] 767455714Skris 767555714Skris *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 767655714Skris yet... 767755714Skris [Steve Henson] 767855714Skris 767955714Skris *) New variables $(RANLIB) and $(PERL) in the Makefiles. 768055714Skris [Ulf M�ller] 768155714Skris 768255714Skris *) New config option to avoid instructions that are illegal on the 80386. 768355714Skris The default code is faster, but requires at least a 486. 768455714Skris [Ulf M�ller] 768555714Skris 768655714Skris *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 768755714Skris SSL2_SERVER_VERSION (not used at all) macros, which are now the 768855714Skris same as SSL2_VERSION anyway. 768955714Skris [Bodo Moeller] 769055714Skris 769155714Skris *) New "-showcerts" option for s_client. 769255714Skris [Bodo Moeller] 769355714Skris 769455714Skris *) Still more PKCS#12 integration. Add pkcs12 application to openssl 769555714Skris application. Various cleanups and fixes. 769655714Skris [Steve Henson] 769755714Skris 769855714Skris *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 769955714Skris modify error routines to work internally. Add error codes and PBE init 770055714Skris to library startup routines. 770155714Skris [Steve Henson] 770255714Skris 770355714Skris *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 770455714Skris packing functions to asn1 and evp. Changed function names and error 770555714Skris codes along the way. 770655714Skris [Steve Henson] 770755714Skris 770855714Skris *) PKCS12 integration: and so it begins... First of several patches to 770955714Skris slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 771055714Skris objects to objects.h 771155714Skris [Steve Henson] 771255714Skris 771355714Skris *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 771455714Skris and display support for Thawte strong extranet extension. 771555714Skris [Steve Henson] 771655714Skris 771755714Skris *) Add LinuxPPC support. 771855714Skris [Jeff Dubrule <igor@pobox.org>] 771955714Skris 772055714Skris *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 772155714Skris bn_div_words in alpha.s. 772255714Skris [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 772355714Skris 772455714Skris *) Make sure the RSA OAEP test is skipped under -DRSAref because 772555714Skris OAEP isn't supported when OpenSSL is built with RSAref. 772655714Skris [Ulf Moeller <ulf@fitug.de>] 772755714Skris 772855714Skris *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 772955714Skris so they no longer are missing under -DNOPROTO. 773055714Skris [Soren S. Jorvang <soren@t.dk>] 773155714Skris 773255714Skris 773355714Skris Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 773455714Skris 773555714Skris *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 773655714Skris doesn't work when the session is reused. Coming soon! 773755714Skris [Ben Laurie] 773855714Skris 773955714Skris *) Fix a security hole, that allows sessions to be reused in the wrong 774055714Skris context thus bypassing client cert protection! All software that uses 774155714Skris client certs and session caches in multiple contexts NEEDS PATCHING to 774255714Skris allow session reuse! A fuller solution is in the works. 774355714Skris [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 774455714Skris 774555714Skris *) Some more source tree cleanups (removed obsolete files 774655714Skris crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 774755714Skris permission on "config" script to be executable) and a fix for the INSTALL 774855714Skris document. 774955714Skris [Ulf Moeller <ulf@fitug.de>] 775055714Skris 775155714Skris *) Remove some legacy and erroneous uses of malloc, free instead of 775255714Skris Malloc, Free. 775355714Skris [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 775455714Skris 775555714Skris *) Make rsa_oaep_test return non-zero on error. 775655714Skris [Ulf Moeller <ulf@fitug.de>] 775755714Skris 775855714Skris *) Add support for native Solaris shared libraries. Configure 775955714Skris solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 776055714Skris if someone would make that last step automatic. 776155714Skris [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 776255714Skris 776355714Skris *) ctx_size was not built with the right compiler during "make links". Fixed. 776455714Skris [Ben Laurie] 776555714Skris 776655714Skris *) Change the meaning of 'ALL' in the cipher list. It now means "everything 776755714Skris except NULL ciphers". This means the default cipher list will no longer 776855714Skris enable NULL ciphers. They need to be specifically enabled e.g. with 776955714Skris the string "DEFAULT:eNULL". 777055714Skris [Steve Henson] 777155714Skris 777255714Skris *) Fix to RSA private encryption routines: if p < q then it would 777355714Skris occasionally produce an invalid result. This will only happen with 777455714Skris externally generated keys because OpenSSL (and SSLeay) ensure p > q. 777555714Skris [Steve Henson] 777655714Skris 777755714Skris *) Be less restrictive and allow also `perl util/perlpath.pl 777855714Skris /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 777955714Skris because this way one can also use an interpreter named `perl5' (which is 778055714Skris usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 778155714Skris installed as `perl'). 778255714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 778355714Skris 778455714Skris *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 778555714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 778655714Skris 778755714Skris *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 778855714Skris advapi32.lib to Win32 build and change the pem test comparision 778955714Skris to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 779055714Skris suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 779155714Skris and crypto/des/ede_cbcm_enc.c. 779255714Skris [Steve Henson] 779355714Skris 779455714Skris *) DES quad checksum was broken on big-endian architectures. Fixed. 779555714Skris [Ben Laurie] 779655714Skris 779755714Skris *) Comment out two functions in bio.h that aren't implemented. Fix up the 779855714Skris Win32 test batch file so it (might) work again. The Win32 test batch file 779955714Skris is horrible: I feel ill.... 780055714Skris [Steve Henson] 780155714Skris 780255714Skris *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 780355714Skris in e_os.h. Audit of header files to check ANSI and non ANSI 780455714Skris sections: 10 functions were absent from non ANSI section and not exported 780555714Skris from Windows DLLs. Fixed up libeay.num for new functions. 780655714Skris [Steve Henson] 780755714Skris 780855714Skris *) Make `openssl version' output lines consistent. 780955714Skris [Ralf S. Engelschall] 781055714Skris 781155714Skris *) Fix Win32 symbol export lists for BIO functions: Added 781255714Skris BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 781355714Skris to ms/libeay{16,32}.def. 781455714Skris [Ralf S. Engelschall] 781555714Skris 781655714Skris *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 781755714Skris fine under Unix and passes some trivial tests I've now added. But the 781855714Skris whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 781955714Skris added to make sure no one expects that this stuff really works in the 782055714Skris OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 782155714Skris up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 782255714Skris openssl_bio.xs. 782355714Skris [Ralf S. Engelschall] 782455714Skris 782555714Skris *) Fix the generation of two part addresses in perl. 782655714Skris [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 782755714Skris 782855714Skris *) Add config entry for Linux on MIPS. 782955714Skris [John Tobey <jtobey@channel1.com>] 783055714Skris 783155714Skris *) Make links whenever Configure is run, unless we are on Windoze. 783255714Skris [Ben Laurie] 783355714Skris 783455714Skris *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 783555714Skris Currently only issuerAltName and AuthorityKeyIdentifier make any sense 783655714Skris in CRLs. 783755714Skris [Steve Henson] 783855714Skris 783955714Skris *) Add a useful kludge to allow package maintainers to specify compiler and 784055714Skris other platforms details on the command line without having to patch the 784155714Skris Configure script everytime: One now can use ``perl Configure 784255714Skris <id>:<details>'', i.e. platform ids are allowed to have details appended 784355714Skris to them (seperated by colons). This is treated as there would be a static 784455714Skris pre-configured entry in Configure's %table under key <id> with value 784555714Skris <details> and ``perl Configure <id>'' is called. So, when you want to 784655714Skris perform a quick test-compile under FreeBSD 3.1 with pgcc and without 784755714Skris assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 784855714Skris now, which overrides the FreeBSD-elf entry on-the-fly. 784955714Skris [Ralf S. Engelschall] 785055714Skris 785155714Skris *) Disable new TLS1 ciphersuites by default: they aren't official yet. 785255714Skris [Ben Laurie] 785355714Skris 785455714Skris *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 785555714Skris on the `perl Configure ...' command line. This way one can compile 785655714Skris OpenSSL libraries with Position Independent Code (PIC) which is needed 785755714Skris for linking it into DSOs. 785855714Skris [Ralf S. Engelschall] 785955714Skris 786055714Skris *) Remarkably, export ciphers were totally broken and no-one had noticed! 786155714Skris Fixed. 786255714Skris [Ben Laurie] 786355714Skris 786455714Skris *) Cleaned up the LICENSE document: The official contact for any license 786555714Skris questions now is the OpenSSL core team under openssl-core@openssl.org. 786655714Skris And add a paragraph about the dual-license situation to make sure people 786755714Skris recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 786855714Skris to the OpenSSL toolkit. 786955714Skris [Ralf S. Engelschall] 787055714Skris 787155714Skris *) General source tree makefile cleanups: Made `making xxx in yyy...' 787255714Skris display consistent in the source tree and replaced `/bin/rm' by `rm'. 787355714Skris Additonally cleaned up the `make links' target: Remove unnecessary 787455714Skris semicolons, subsequent redundant removes, inline point.sh into mklink.sh 787555714Skris to speed processing and no longer clutter the display with confusing 787655714Skris stuff. Instead only the actually done links are displayed. 787755714Skris [Ralf S. Engelschall] 787855714Skris 787955714Skris *) Permit null encryption ciphersuites, used for authentication only. It used 788055714Skris to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 788155714Skris It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 788255714Skris encryption. 788355714Skris [Ben Laurie] 788455714Skris 788555714Skris *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 788655714Skris signed attributes when verifying signatures (this would break them), 788755714Skris the detached data encoding was wrong and public keys obtained using 788855714Skris X509_get_pubkey() weren't freed. 788955714Skris [Steve Henson] 789055714Skris 789155714Skris *) Add text documentation for the BUFFER functions. Also added a work around 789255714Skris to a Win95 console bug. This was triggered by the password read stuff: the 789355714Skris last character typed gets carried over to the next fread(). If you were 789455714Skris generating a new cert request using 'req' for example then the last 789555714Skris character of the passphrase would be CR which would then enter the first 789655714Skris field as blank. 789755714Skris [Steve Henson] 789855714Skris 789955714Skris *) Added the new `Includes OpenSSL Cryptography Software' button as 790055714Skris doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 790155714Skris button and can be used by applications based on OpenSSL to show the 790255714Skris relationship to the OpenSSL project. 790355714Skris [Ralf S. Engelschall] 790455714Skris 790555714Skris *) Remove confusing variables in function signatures in files 790655714Skris ssl/ssl_lib.c and ssl/ssl.h. 790755714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 790855714Skris 790955714Skris *) Don't install bss_file.c under PREFIX/include/ 791055714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 791155714Skris 791255714Skris *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 791355714Skris functions that return function pointers and has support for NT specific 791455714Skris stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 791555714Skris #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 791655714Skris unsigned to signed types: this was killing the Win32 compile. 791755714Skris [Steve Henson] 791855714Skris 791955714Skris *) Add new certificate file to stack functions, 792055714Skris SSL_add_dir_cert_subjects_to_stack() and 792155714Skris SSL_add_file_cert_subjects_to_stack(). These largely supplant 792255714Skris SSL_load_client_CA_file(), and can be used to add multiple certs easily 792355714Skris to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 792455714Skris This means that Apache-SSL and similar packages don't have to mess around 792555714Skris to add as many CAs as they want to the preferred list. 792655714Skris [Ben Laurie] 792755714Skris 792855714Skris *) Experiment with doxygen documentation. Currently only partially applied to 792955714Skris ssl/ssl_lib.c. 793055714Skris See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 793155714Skris openssl.doxy as the configuration file. 793255714Skris [Ben Laurie] 793355714Skris 793455714Skris *) Get rid of remaining C++-style comments which strict C compilers hate. 793555714Skris [Ralf S. Engelschall, pointed out by Carlos Amengual] 793655714Skris 793755714Skris *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 793855714Skris compiled in by default: it has problems with large keys. 793955714Skris [Steve Henson] 794055714Skris 794155714Skris *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 794255714Skris DH private keys and/or callback functions which directly correspond to 794355714Skris their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 794455714Skris is needed for applications which have to configure certificates on a 794555714Skris per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 794655714Skris (e.g. s_server). 794755714Skris For the RSA certificate situation is makes no difference, but 794855714Skris for the DSA certificate situation this fixes the "no shared cipher" 794955714Skris problem where the OpenSSL cipher selection procedure failed because the 795055714Skris temporary keys were not overtaken from the context and the API provided 795155714Skris no way to reconfigure them. 795255714Skris The new functions now let applications reconfigure the stuff and they 795355714Skris are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 795455714Skris SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 795555714Skris non-public-API function ssl_cert_instantiate() is used as a helper 795655714Skris function and also to reduce code redundancy inside ssl_rsa.c. 795755714Skris [Ralf S. Engelschall] 795855714Skris 795955714Skris *) Move s_server -dcert and -dkey options out of the undocumented feature 796055714Skris area because they are useful for the DSA situation and should be 796155714Skris recognized by the users. 796255714Skris [Ralf S. Engelschall] 796355714Skris 796455714Skris *) Fix the cipher decision scheme for export ciphers: the export bits are 796555714Skris *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 796655714Skris SSL_EXP_MASK. So, the original variable has to be used instead of the 796755714Skris already masked variable. 796855714Skris [Richard Levitte <levitte@stacken.kth.se>] 796955714Skris 797055714Skris *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 797155714Skris [Richard Levitte <levitte@stacken.kth.se>] 797255714Skris 797355714Skris *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 797455714Skris from `int' to `unsigned int' because it's a length and initialized by 797555714Skris EVP_DigestFinal() which expects an `unsigned int *'. 797655714Skris [Richard Levitte <levitte@stacken.kth.se>] 797755714Skris 797855714Skris *) Don't hard-code path to Perl interpreter on shebang line of Configure 797955714Skris script. Instead use the usual Shell->Perl transition trick. 798055714Skris [Ralf S. Engelschall] 798155714Skris 798255714Skris *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 798355714Skris (in addition to RSA certificates) to match the behaviour of `openssl dsa 798455714Skris -noout -modulus' as it's already the case for `openssl rsa -noout 798555714Skris -modulus'. For RSA the -modulus is the real "modulus" while for DSA 798655714Skris currently the public key is printed (a decision which was already done by 798755714Skris `openssl dsa -modulus' in the past) which serves a similar purpose. 798855714Skris Additionally the NO_RSA no longer completely removes the whole -modulus 798955714Skris option; it now only avoids using the RSA stuff. Same applies to NO_DSA 799055714Skris now, too. 799155714Skris [Ralf S. Engelschall] 799255714Skris 799355714Skris *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 799455714Skris BIO. See the source (crypto/evp/bio_ok.c) for more info. 799555714Skris [Arne Ansper <arne@ats.cyber.ee>] 799655714Skris 799755714Skris *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 799855714Skris to be added. Now both 'req' and 'ca' can use new objects defined in the 799955714Skris config file. 800055714Skris [Steve Henson] 800155714Skris 800255714Skris *) Add cool BIO that does syslog (or event log on NT). 800355714Skris [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 800455714Skris 800555714Skris *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 800655714Skris TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 800755714Skris TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 800855714Skris Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 800955714Skris [Ben Laurie] 801055714Skris 801155714Skris *) Add preliminary config info for new extension code. 801255714Skris [Steve Henson] 801355714Skris 801455714Skris *) Make RSA_NO_PADDING really use no padding. 801555714Skris [Ulf Moeller <ulf@fitug.de>] 801655714Skris 801755714Skris *) Generate errors when private/public key check is done. 801855714Skris [Ben Laurie] 801955714Skris 802055714Skris *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 802155714Skris for some CRL extensions and new objects added. 802255714Skris [Steve Henson] 802355714Skris 802455714Skris *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 802555714Skris key usage extension and fuller support for authority key id. 802655714Skris [Steve Henson] 802755714Skris 802855714Skris *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 802955714Skris padding method for RSA, which is recommended for new applications in PKCS 803055714Skris #1 v2.0 (RFC 2437, October 1998). 803155714Skris OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 803255714Skris foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 803355714Skris against Bleichbacher's attack on RSA. 803455714Skris [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 803555714Skris Ben Laurie] 803655714Skris 803755714Skris *) Updates to the new SSL compression code 803855714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 803955714Skris 804055714Skris *) Fix so that the version number in the master secret, when passed 804155714Skris via RSA, checks that if TLS was proposed, but we roll back to SSLv3 804255714Skris (because the server will not accept higher), that the version number 804355714Skris is 0x03,0x01, not 0x03,0x00 804455714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 804555714Skris 804655714Skris *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 804755714Skris leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 804855714Skris in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 804955714Skris [Steve Henson] 805055714Skris 805155714Skris *) Support for RAW extensions where an arbitrary extension can be 805255714Skris created by including its DER encoding. See apps/openssl.cnf for 805355714Skris an example. 805455714Skris [Steve Henson] 805555714Skris 805655714Skris *) Make sure latest Perl versions don't interpret some generated C array 805755714Skris code as Perl array code in the crypto/err/err_genc.pl script. 805855714Skris [Lars Weber <3weber@informatik.uni-hamburg.de>] 805955714Skris 806055714Skris *) Modify ms/do_ms.bat to not generate assembly language makefiles since 806155714Skris not many people have the assembler. Various Win32 compilation fixes and 806255714Skris update to the INSTALL.W32 file with (hopefully) more accurate Win32 806355714Skris build instructions. 806455714Skris [Steve Henson] 806555714Skris 806655714Skris *) Modify configure script 'Configure' to automatically create crypto/date.h 806755714Skris file under Win32 and also build pem.h from pem.org. New script 806855714Skris util/mkfiles.pl to create the MINFO file on environments that can't do a 806955714Skris 'make files': perl util/mkfiles.pl >MINFO should work. 807055714Skris [Steve Henson] 807155714Skris 807255714Skris *) Major rework of DES function declarations, in the pursuit of correctness 807355714Skris and purity. As a result, many evil casts evaporated, and some weirdness, 807455714Skris too. You may find this causes warnings in your code. Zapping your evil 807555714Skris casts will probably fix them. Mostly. 807655714Skris [Ben Laurie] 807755714Skris 807855714Skris *) Fix for a typo in asn1.h. Bug fix to object creation script 807955714Skris obj_dat.pl. It considered a zero in an object definition to mean 808055714Skris "end of object": none of the objects in objects.h have any zeros 808155714Skris so it wasn't spotted. 808255714Skris [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 808355714Skris 808455714Skris *) Add support for Triple DES Cipher Block Chaining with Output Feedback 808555714Skris Masking (CBCM). In the absence of test vectors, the best I have been able 808655714Skris to do is check that the decrypt undoes the encrypt, so far. Send me test 808755714Skris vectors if you have them. 808855714Skris [Ben Laurie] 808955714Skris 809055714Skris *) Correct calculation of key length for export ciphers (too much space was 809155714Skris allocated for null ciphers). This has not been tested! 809255714Skris [Ben Laurie] 809355714Skris 809455714Skris *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 809555714Skris message is now correct (it understands "crypto" and "ssl" on its 809655714Skris command line). There is also now an "update" option. This will update 809755714Skris the util/ssleay.num and util/libeay.num files with any new functions. 809855714Skris If you do a: 809955714Skris perl util/mkdef.pl crypto ssl update 810055714Skris it will update them. 810155714Skris [Steve Henson] 810255714Skris 810355714Skris *) Overhauled the Perl interface (perl/*): 810455714Skris - ported BN stuff to OpenSSL's different BN library 810555714Skris - made the perl/ source tree CVS-aware 810655714Skris - renamed the package from SSLeay to OpenSSL (the files still contain 810755714Skris their history because I've copied them in the repository) 810855714Skris - removed obsolete files (the test scripts will be replaced 810955714Skris by better Test::Harness variants in the future) 811055714Skris [Ralf S. Engelschall] 811155714Skris 811255714Skris *) First cut for a very conservative source tree cleanup: 811355714Skris 1. merge various obsolete readme texts into doc/ssleay.txt 811455714Skris where we collect the old documents and readme texts. 811555714Skris 2. remove the first part of files where I'm already sure that we no 811655714Skris longer need them because of three reasons: either they are just temporary 811755714Skris files which were left by Eric or they are preserved original files where 811855714Skris I've verified that the diff is also available in the CVS via "cvs diff 811955714Skris -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 812055714Skris the crypto/md/ stuff). 812155714Skris [Ralf S. Engelschall] 812255714Skris 812355714Skris *) More extension code. Incomplete support for subject and issuer alt 812455714Skris name, issuer and authority key id. Change the i2v function parameters 812555714Skris and add an extra 'crl' parameter in the X509V3_CTX structure: guess 812655714Skris what that's for :-) Fix to ASN1 macro which messed up 812755714Skris IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 812855714Skris [Steve Henson] 812955714Skris 813055714Skris *) Preliminary support for ENUMERATED type. This is largely copied from the 813155714Skris INTEGER code. 813255714Skris [Steve Henson] 813355714Skris 813455714Skris *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 813555714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 813655714Skris 813755714Skris *) Make sure `make rehash' target really finds the `openssl' program. 813855714Skris [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 813955714Skris 814055714Skris *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 814155714Skris like to hear about it if this slows down other processors. 814255714Skris [Ben Laurie] 814355714Skris 814455714Skris *) Add CygWin32 platform information to Configure script. 814555714Skris [Alan Batie <batie@aahz.jf.intel.com>] 814655714Skris 814755714Skris *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 814855714Skris [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 814955714Skris 815055714Skris *) New program nseq to manipulate netscape certificate sequences 815155714Skris [Steve Henson] 815255714Skris 815355714Skris *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 815455714Skris few typos. 815555714Skris [Steve Henson] 815655714Skris 815755714Skris *) Fixes to BN code. Previously the default was to define BN_RECURSION 815855714Skris but the BN code had some problems that would cause failures when 815955714Skris doing certificate verification and some other functions. 816055714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 816155714Skris 816255714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 816355714Skris [Steve Henson] 816455714Skris 816555714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 816655714Skris [Steve Henson] 816755714Skris 816855714Skris *) Add several PKIX and private extended key usage OIDs. 816955714Skris [Steve Henson] 817055714Skris 817155714Skris *) Modify the 'ca' program to handle the new extension code. Modify 817255714Skris openssl.cnf for new extension format, add comments. 817355714Skris [Steve Henson] 817455714Skris 817555714Skris *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 817655714Skris and add a sample to openssl.cnf so req -x509 now adds appropriate 817755714Skris CA extensions. 817855714Skris [Steve Henson] 817955714Skris 818055714Skris *) Continued X509 V3 changes. Add to other makefiles, integrate with the 818155714Skris error code, add initial support to X509_print() and x509 application. 818255714Skris [Steve Henson] 818355714Skris 818455714Skris *) Takes a deep breath and start addding X509 V3 extension support code. Add 818555714Skris files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 818655714Skris stuff is currently isolated and isn't even compiled yet. 818755714Skris [Steve Henson] 818855714Skris 818955714Skris *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 819055714Skris ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 819155714Skris Removed the versions check from X509 routines when loading extensions: 819255714Skris this allows certain broken certificates that don't set the version 819355714Skris properly to be processed. 819455714Skris [Steve Henson] 819555714Skris 819655714Skris *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 819755714Skris Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 819855714Skris can still be regenerated with "make depend". 819955714Skris [Ben Laurie] 820055714Skris 820155714Skris *) Spelling mistake in C version of CAST-128. 820255714Skris [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 820355714Skris 820455714Skris *) Changes to the error generation code. The perl script err-code.pl 820555714Skris now reads in the old error codes and retains the old numbers, only 820655714Skris adding new ones if necessary. It also only changes the .err files if new 820755714Skris codes are added. The makefiles have been modified to only insert errors 820855714Skris when needed (to avoid needlessly modifying header files). This is done 820955714Skris by only inserting errors if the .err file is newer than the auto generated 821055714Skris C file. To rebuild all the error codes from scratch (the old behaviour) 821155714Skris either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 821255714Skris or delete all the .err files. 821355714Skris [Steve Henson] 821455714Skris 821555714Skris *) CAST-128 was incorrectly implemented for short keys. The C version has 821655714Skris been fixed, but is untested. The assembler versions are also fixed, but 821755714Skris new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 821855714Skris to regenerate it if needed. 821955714Skris [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 822055714Skris Hagino <itojun@kame.net>] 822155714Skris 822255714Skris *) File was opened incorrectly in randfile.c. 822355714Skris [Ulf M�ller <ulf@fitug.de>] 822455714Skris 822555714Skris *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 822655714Skris functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 822755714Skris GeneralizedTime. ASN1_TIME is the proper type used in certificates et 822855714Skris al: it's just almost always a UTCTime. Note this patch adds new error 822955714Skris codes so do a "make errors" if there are problems. 823055714Skris [Steve Henson] 823155714Skris 823255714Skris *) Correct Linux 1 recognition in config. 823355714Skris [Ulf M�ller <ulf@fitug.de>] 823455714Skris 823555714Skris *) Remove pointless MD5 hash when using DSA keys in ca. 823655714Skris [Anonymous <nobody@replay.com>] 823755714Skris 823855714Skris *) Generate an error if given an empty string as a cert directory. Also 823955714Skris generate an error if handed NULL (previously returned 0 to indicate an 824055714Skris error, but didn't set one). 824155714Skris [Ben Laurie, reported by Anonymous <nobody@replay.com>] 824255714Skris 824355714Skris *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 824455714Skris [Ben Laurie] 824555714Skris 824655714Skris *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 824755714Skris parameters. This was causing a warning which killed off the Win32 compile. 824855714Skris [Steve Henson] 824955714Skris 825055714Skris *) Remove C++ style comments from crypto/bn/bn_local.h. 825155714Skris [Neil Costigan <neil.costigan@celocom.com>] 825255714Skris 825355714Skris *) The function OBJ_txt2nid was broken. It was supposed to return a nid 825455714Skris based on a text string, looking up short and long names and finally 825555714Skris "dot" format. The "dot" format stuff didn't work. Added new function 825655714Skris OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 825755714Skris OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 825855714Skris OID is not part of the table. 825955714Skris [Steve Henson] 826055714Skris 826155714Skris *) Add prototypes to X509 lookup/verify methods, fixing a bug in 826255714Skris X509_LOOKUP_by_alias(). 826355714Skris [Ben Laurie] 826455714Skris 826555714Skris *) Sort openssl functions by name. 826655714Skris [Ben Laurie] 826755714Skris 826855714Skris *) Get the gendsa program working (hopefully) and add it to app list. Remove 826955714Skris encryption from sample DSA keys (in case anyone is interested the password 827055714Skris was "1234"). 827155714Skris [Steve Henson] 827255714Skris 827355714Skris *) Make _all_ *_free functions accept a NULL pointer. 827455714Skris [Frans Heymans <fheymans@isaserver.be>] 827555714Skris 827655714Skris *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 827755714Skris NULL pointers. 827855714Skris [Anonymous <nobody@replay.com>] 827955714Skris 828055714Skris *) s_server should send the CAfile as acceptable CAs, not its own cert. 828155714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 828255714Skris 828355714Skris *) Don't blow it for numeric -newkey arguments to apps/req. 828455714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 828555714Skris 828655714Skris *) Temp key "for export" tests were wrong in s3_srvr.c. 828755714Skris [Anonymous <nobody@replay.com>] 828855714Skris 828955714Skris *) Add prototype for temp key callback functions 829055714Skris SSL_CTX_set_tmp_{rsa,dh}_callback(). 829155714Skris [Ben Laurie] 829255714Skris 829355714Skris *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 829455714Skris DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 829555714Skris [Steve Henson] 829655714Skris 829755714Skris *) X509_name_add_entry() freed the wrong thing after an error. 829855714Skris [Arne Ansper <arne@ats.cyber.ee>] 829955714Skris 830055714Skris *) rsa_eay.c would attempt to free a NULL context. 830155714Skris [Arne Ansper <arne@ats.cyber.ee>] 830255714Skris 830355714Skris *) BIO_s_socket() had a broken should_retry() on Windoze. 830455714Skris [Arne Ansper <arne@ats.cyber.ee>] 830555714Skris 830655714Skris *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 830755714Skris [Arne Ansper <arne@ats.cyber.ee>] 830855714Skris 830955714Skris *) Make sure the already existing X509_STORE->depth variable is initialized 831055714Skris in X509_STORE_new(), but document the fact that this variable is still 831155714Skris unused in the certificate verification process. 831255714Skris [Ralf S. Engelschall] 831355714Skris 831455714Skris *) Fix the various library and apps files to free up pkeys obtained from 831555714Skris X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 831655714Skris [Steve Henson] 831755714Skris 831855714Skris *) Fix reference counting in X509_PUBKEY_get(). This makes 831955714Skris demos/maurice/example2.c work, amongst others, probably. 832055714Skris [Steve Henson and Ben Laurie] 832155714Skris 832255714Skris *) First cut of a cleanup for apps/. First the `ssleay' program is now named 832355714Skris `openssl' and second, the shortcut symlinks for the `openssl <command>' 832455714Skris are no longer created. This way we have a single and consistent command 832555714Skris line interface `openssl <command>', similar to `cvs <command>'. 832655714Skris [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 832755714Skris 832855714Skris *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 832955714Skris BIT STRING wrapper always have zero unused bits. 833055714Skris [Steve Henson] 833155714Skris 833255714Skris *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 833355714Skris [Steve Henson] 833455714Skris 833555714Skris *) Make the top-level INSTALL documentation easier to understand. 833655714Skris [Paul Sutton] 833755714Skris 833855714Skris *) Makefiles updated to exit if an error occurs in a sub-directory 833955714Skris make (including if user presses ^C) [Paul Sutton] 834055714Skris 834155714Skris *) Make Montgomery context stuff explicit in RSA data structure. 834255714Skris [Ben Laurie] 834355714Skris 834455714Skris *) Fix build order of pem and err to allow for generated pem.h. 834555714Skris [Ben Laurie] 834655714Skris 834755714Skris *) Fix renumbering bug in X509_NAME_delete_entry(). 834855714Skris [Ben Laurie] 834955714Skris 835055714Skris *) Enhanced the err-ins.pl script so it makes the error library number 835155714Skris global and can add a library name. This is needed for external ASN1 and 835255714Skris other error libraries. 835355714Skris [Steve Henson] 835455714Skris 835555714Skris *) Fixed sk_insert which never worked properly. 835655714Skris [Steve Henson] 835755714Skris 835855714Skris *) Fix ASN1 macros so they can handle indefinite length construted 835955714Skris EXPLICIT tags. Some non standard certificates use these: they can now 836055714Skris be read in. 836155714Skris [Steve Henson] 836255714Skris 836355714Skris *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 836455714Skris into a single doc/ssleay.txt bundle. This way the information is still 836555714Skris preserved but no longer messes up this directory. Now it's new room for 836655714Skris the new set of documenation files. 836755714Skris [Ralf S. Engelschall] 836855714Skris 836955714Skris *) SETs were incorrectly DER encoded. This was a major pain, because they 837055714Skris shared code with SEQUENCEs, which aren't coded the same. This means that 837155714Skris almost everything to do with SETs or SEQUENCEs has either changed name or 837255714Skris number of arguments. 837355714Skris [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 837455714Skris 837555714Skris *) Fix test data to work with the above. 837655714Skris [Ben Laurie] 837755714Skris 837855714Skris *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 837955714Skris was already fixed by Eric for 0.9.1 it seems. 838055714Skris [Ben Laurie - pointed out by Ulf M�ller <ulf@fitug.de>] 838155714Skris 838255714Skris *) Autodetect FreeBSD3. 838355714Skris [Ben Laurie] 838455714Skris 838555714Skris *) Fix various bugs in Configure. This affects the following platforms: 838655714Skris nextstep 838755714Skris ncr-scde 838855714Skris unixware-2.0 838955714Skris unixware-2.0-pentium 839055714Skris sco5-cc. 839155714Skris [Ben Laurie] 839255714Skris 839355714Skris *) Eliminate generated files from CVS. Reorder tests to regenerate files 839455714Skris before they are needed. 839555714Skris [Ben Laurie] 839655714Skris 839755714Skris *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 839855714Skris [Ben Laurie] 839955714Skris 840055714Skris 840155714Skris Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 840255714Skris 840355714Skris *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 840455714Skris changed SSLeay to OpenSSL in version strings. 840555714Skris [Ralf S. Engelschall] 840655714Skris 840755714Skris *) Some fixups to the top-level documents. 840855714Skris [Paul Sutton] 840955714Skris 841055714Skris *) Fixed the nasty bug where rsaref.h was not found under compile-time 841155714Skris because the symlink to include/ was missing. 841255714Skris [Ralf S. Engelschall] 841355714Skris 841455714Skris *) Incorporated the popular no-RSA/DSA-only patches 841555714Skris which allow to compile a RSA-free SSLeay. 841655714Skris [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 841755714Skris 841855714Skris *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 841955714Skris when "ssleay" is still not found. 842055714Skris [Ralf S. Engelschall] 842155714Skris 842255714Skris *) Added more platforms to Configure: Cray T3E, HPUX 11, 842355714Skris [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 842455714Skris 842555714Skris *) Updated the README file. 842655714Skris [Ralf S. Engelschall] 842755714Skris 842855714Skris *) Added various .cvsignore files in the CVS repository subdirs 842955714Skris to make a "cvs update" really silent. 843055714Skris [Ralf S. Engelschall] 843155714Skris 843255714Skris *) Recompiled the error-definition header files and added 843355714Skris missing symbols to the Win32 linker tables. 843455714Skris [Ralf S. Engelschall] 843555714Skris 843655714Skris *) Cleaned up the top-level documents; 843755714Skris o new files: CHANGES and LICENSE 843855714Skris o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 843955714Skris o merged COPYRIGHT into LICENSE 844055714Skris o removed obsolete TODO file 844155714Skris o renamed MICROSOFT to INSTALL.W32 844255714Skris [Ralf S. Engelschall] 844355714Skris 844455714Skris *) Removed dummy files from the 0.9.1b source tree: 844555714Skris crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 844655714Skris crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 844755714Skris crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 844855714Skris crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 844955714Skris util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 845055714Skris [Ralf S. Engelschall] 845155714Skris 845255714Skris *) Added various platform portability fixes. 845355714Skris [Mark J. Cox] 845455714Skris 845555714Skris *) The Genesis of the OpenSSL rpject: 845655714Skris We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 845755714Skris Young and Tim J. Hudson created while they were working for C2Net until 845855714Skris summer 1998. 845955714Skris [The OpenSSL Project] 846055714Skris 846155714Skris 846255714Skris Changes between 0.9.0b and 0.9.1b [not released] 846355714Skris 846455714Skris *) Updated a few CA certificates under certs/ 846555714Skris [Eric A. Young] 846655714Skris 846755714Skris *) Changed some BIGNUM api stuff. 846855714Skris [Eric A. Young] 846955714Skris 847055714Skris *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 847155714Skris DGUX x86, Linux Alpha, etc. 847255714Skris [Eric A. Young] 847355714Skris 847455714Skris *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 847555714Skris RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 847655714Skris available). 847755714Skris [Eric A. Young] 847855714Skris 847955714Skris *) Add -strparse option to asn1pars program which parses nested 848055714Skris binary structures 848155714Skris [Dr Stephen Henson <shenson@bigfoot.com>] 848255714Skris 848355714Skris *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 848455714Skris [Eric A. Young] 848555714Skris 848655714Skris *) DSA fix for "ca" program. 848755714Skris [Eric A. Young] 848855714Skris 848955714Skris *) Added "-genkey" option to "dsaparam" program. 849055714Skris [Eric A. Young] 849155714Skris 849255714Skris *) Added RIPE MD160 (rmd160) message digest. 849355714Skris [Eric A. Young] 849455714Skris 849555714Skris *) Added -a (all) option to "ssleay version" command. 849655714Skris [Eric A. Young] 849755714Skris 849855714Skris *) Added PLATFORM define which is the id given to Configure. 849955714Skris [Eric A. Young] 850055714Skris 850155714Skris *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 850255714Skris [Eric A. Young] 850355714Skris 850455714Skris *) Extended the ASN.1 parser routines. 850555714Skris [Eric A. Young] 850655714Skris 850755714Skris *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 850855714Skris [Eric A. Young] 850955714Skris 851055714Skris *) Added a BN_CTX to the BN library. 851155714Skris [Eric A. Young] 851255714Skris 851355714Skris *) Fixed the weak key values in DES library 851455714Skris [Eric A. Young] 851555714Skris 851655714Skris *) Changed API in EVP library for cipher aliases. 851755714Skris [Eric A. Young] 851855714Skris 851955714Skris *) Added support for RC2/64bit cipher. 852055714Skris [Eric A. Young] 852155714Skris 852255714Skris *) Converted the lhash library to the crypto/mem.c functions. 852355714Skris [Eric A. Young] 852455714Skris 852555714Skris *) Added more recognized ASN.1 object ids. 852655714Skris [Eric A. Young] 852755714Skris 852855714Skris *) Added more RSA padding checks for SSL/TLS. 852955714Skris [Eric A. Young] 853055714Skris 853155714Skris *) Added BIO proxy/filter functionality. 853255714Skris [Eric A. Young] 853355714Skris 853455714Skris *) Added extra_certs to SSL_CTX which can be used 853555714Skris send extra CA certificates to the client in the CA cert chain sending 853655714Skris process. It can be configured with SSL_CTX_add_extra_chain_cert(). 853755714Skris [Eric A. Young] 853855714Skris 853955714Skris *) Now Fortezza is denied in the authentication phase because 854055714Skris this is key exchange mechanism is not supported by SSLeay at all. 854155714Skris [Eric A. Young] 854255714Skris 854355714Skris *) Additional PKCS1 checks. 854455714Skris [Eric A. Young] 854555714Skris 854655714Skris *) Support the string "TLSv1" for all TLS v1 ciphers. 854755714Skris [Eric A. Young] 854855714Skris 854955714Skris *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 855055714Skris ex_data index of the SSL context in the X509_STORE_CTX ex_data. 855155714Skris [Eric A. Young] 855255714Skris 855355714Skris *) Fixed a few memory leaks. 855455714Skris [Eric A. Young] 855555714Skris 855655714Skris *) Fixed various code and comment typos. 855755714Skris [Eric A. Young] 855855714Skris 855955714Skris *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 856055714Skris bytes sent in the client random. 856155714Skris [Edward Bishop <ebishop@spyglass.com>] 856255714Skris 8563